Stuart D. Gathman wrote:
compilers, on the other hand, would read the source records -
and source records would include text files fetched via HTTP
and v=spf1 records.
The more I think about it, the more I like that idea. So
shoot holes in it quick before I start promoting it.
For the DynDNS cases you wouldn't want to "compile" A whenever
its TTL expires, it's bad enough to fetch it when a checker
really needs it. The TTL of a "compiled" record would be the
worst case in its sources.
It sounds nice at first glance, but it's actually a bad idea
for e.g. include with tons of "a include:isp.example -all" and
different domains (different As), why "compile" that again and
again with the same convoluted isp.example policy ?
Include and redirect are also different ways to split a long
policy into different "source lines", if you've "compiled" it
you still need a way to split very long results into chunks
working with UDP.
Mechanisms like exists are probably beyond any "compilation".
There are also subtle points with the semantics, it's different
for IPv4 vs. IPv6, you'd need two "compilations".
Frank, dreaming of a v=spf1.001 standard :-)
-------
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