In <42DD6D96(_dot_)6650(_at_)xyzzy(_dot_)claranet(_dot_)de> Frank Ellermann
<nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> writes:
JFTR, when Wayne talked about a time machine back to 2003 it's
more like April 2005 and -01pre2:
<http://mid.gmane.org/4269C17D(_dot_)39B2(_at_)xyzzy(_dot_)claranet(_dot_)de>
I'm no fan of the "silent" limit, but different ideas for MX
and PTR as before -01pre2 (in my interpretation) was too odd.
I disagree with your interpretation of the -01pre2 spec. To the best
of my knowledge, at no point was the MX or PTR limits defined to
return a PermError.
I confess that the DoS limits are based off of my analysis from back
in early 2004 and what I put into libspf2. I think DoS limits are
critical and therefore justified incompatible changes and I used the
"libspf2" was the "existing practices" that I hung my hat on.
So, you might not have needed to go all the way back to 2003, but you
would needed to have gone back to early 2004.
-wayne