-----Ursprungligt meddelande-----
Från: Meng Weng Wong [mailto:mengwong(_at_)dumbo(_dot_)pobox(_dot_)com]
Skickat: den 23 mars 2004 02:41
...
I'm a little bit bothered by the new paragraph because it tries to
micromanage clients in too much detail. Is there a better way to say:
you have to respect an exp modifier even if it appears after "all"?
I was happy with the original interpretation that the processing terminate when
a mechanism matched. It is not that hard to put the explanation before the
mechanism that will fail. This would also allow for different explanations for
different matches:
v=spf1 exp=user._exp.%{d} -exists:%{l}.%{d} mx exp=_exp.%{d} -all
Will give a different explanation if the user is in the list of users not
permitted to send mail and a default explanation for all other failures.
Another reason for this is that i may want to change a modifier during
processing. If we in the future implement a modifier that alters the behaviour
of the include mechanism, and I want to use that for only one of the included
domains, this will not be possible if we parse all modifiers first an then
start procesing the mechanisms.
v=spf1 useparent=yes include:%{d1r}.frienddomain.com useparent=no
include:mail.myisp.net -all
/Johan