[leafnode-list] Re: Header PATH and FQDN

[clemens fischer <ino-news@xxxxxxxxxxxxxxxxxxxxxxxxx>]

Matthias Andree wrote:

> Am 07.12.2010 22:32, schrieb clemens fischer:
>
>> If the RFC 5537 allows this, then leafnode should allow it, too.  The
>> next question would be how to reliably determine this IP or how to
>> configure it for dial-up hosts, but this shouldn't be a hard problem.
>
> While permitted (note that it should be a domain literal then although
> the ABNF syntax does not enforce that), it is discouraged in section
> 3.1.3 (p.12) of RFC 5536 due to lack of uniqueness.
>
> If the host only has RFC-1918 IPv4 addresses, or only IPv6 addresses,
> what are you going to pick?  There are thousands of computers with
> 192.168.178.X and 192.168.2.X address (anything that is a DHCP client
> to an AVM Fritz!Box or Deutsche Telekom Speedport router), too, and we
> don't need our external IP address for any other purpose.
>
> We might as well create some namespace (a subdomain in this
> case) under leafnode.org and use a host-and-time-based UUID, or
> some-128-bit-hash of UUID to mask the MAC address.

If that is possible:  how about an el cheapo hash of the current time,
the (some) article headers and the authors email appended?  With that
email in angle brackets, we could reuse msgid_hash()@msgid_hash.c (no
extra libs), and then there's the algorithm we use for distributing
received articles in the spool directories (can't remember function nor
file).

Would we need a way to expire such IDs?  Do these FQDNs need to be
pingable or verifyable in any way?


clemens

-- 
_______________________________________________
leafnode-list mailing list
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
https://www.dt.e-technik.uni-dortmund.de/mailman/listinfo/leafnode-list
http://leafnode.sourceforge.net/