On Fri, 13 Jan 2006, Frank Ellermann wrote:
Meng, Mark, me, and others, here on this list, summer 2004.
William's counter-proposal was less optimal and confusing like
hell for non-trivial cases. Greg's macro-idea is a feasible
alternative.
New macro means new record version to support it and means new dns lookup.
I emailed Mark (privately) when scopes where finally introduced and
asked to make changes to redirect and include so that they could be
directed TXT record within the same domain but with different version
(i.e. scope specification within include) which could solve some of
these issues and allow sharing parts of the record without new lookup).
Mark did not like it but I don't remember his exact reasons right away,
can dig it up his response if people are interested.
"scope=-pra include:mydomain.esp.com scope=+pra -all" )
This lets you add and delete scope values, instead of just
setting them.
This is an interesting variant I hadn't thought of before.
The elaborated version of William's idea,
It is one of the variants I posted on marid long ago. I had both
the ones with scope being explicitly defined by positional modifier
and the ones with scopes being added and deleted by such modifier.
IIRC it doesn't work,
and that was the point where "consensus positional" was decreed
by the marid-protocol editor (Mark, seconded by me). JFTR, bye
SPF1 defines all modifiers as non-positional. At the same time I don't
think most implementations actually treat them. More importantly as
soon as you reach -all the rest would be ignored so one just needs
to be careful with what is before first "-all" and make sure it is
all basically "mfrom" scope and so would be used by by non-scope-aware
systems as well. So in practice it can in fact work and bring scoping
on top of existing spf1 syntax. How optimal is that is another issue.
I also had separate proposal later to bring scoping to spf1 by means
of non-positional modifier "sc.<scope>=<data>" where <scope> is actual
scope and <data> has special syntax and can identify it as being same as
existing record or point to another SPF record where scope-aware systems
could find the info, i.e. :
v=spf1 ... -all sc.sender=net sc.pra=spf("spf2.0/pra")
spf2.0/pra ...
For supporting multiple scopes together, it would be sc.<scope1>.<scope2>=
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net
-------
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