spf-discuss
[Top] [All Lists]

Re: Re: New SPFv1 spec: draft-schlitt-spf-classic-02pre2

2005-06-05 19:50:32
In <42A39FEB(_dot_)7EB9(_at_)xyzzy(_dot_)claranet(_dot_)de> Frank Ellermann 
<nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> writes:

wayne wrote:
 
We also decided to change "prefix" to "qualifier".

Watch "prefixes" in s/.refix/qualifier/g => "qualifieres" ;-)

I think I cought that case.  If not, that's what the RFC editor is
for, right?  ;-)


let it sit for a while, and then do another read-over.

Why do you use "manual" references for stuff like US-ASCII
or RMX ?  For the Vixie note it's clear.  Hm, okay, I see,
rfc/bibxml2/reference.ANSI.X3-4.1968.xml has no nice anchor,
and RfC 20 won't be much better, PITA.

I confess, I just wanted to get it done.  Yesterday there wasn't a
problem with not having a normative reference for US-ASCII at all.
;-)

The docname still says "-01", you need "-02".  And a new date.

Please don't forget to add a -02 change log for the IESG^Wthe
readers.  Mention anything you did about the "intended status"
for Ted's [Discuss]^W^WBruce.  

Readers probably like to see that something really happened.

Both of these are on my "todo at the last moment" list, along with
removing the note about the draft not being for IESG review.

The ToC is confused, it adds 12.2 below 13.2.  I vaguely recall
similar problems with NNTP draft -26 and an xml2rfc hot fix.

Ugh.  that *is* strange.  Good catch, thanks.


Add the "if a message is rejected" blurb with the proper DSN
codes to "PermError" (same style as TempError/SoftFail/Fail) -
or do you plan to discuss SPF Council resolutions with me on
public IETF lists in the "last call" ?  Maybe we could use
some PayPal "donate" URLs for this spectacle and then retire.

This was decided by a SPF Council resolution and I feel comfortable
with the way it is.  The mengwong-spf-* drafts say that PermError (ok,
"unknown") MUST be treated the same as None.  Until someone provides a
time machine for me to go back and change things, I will support the
draft as it is.  


You forgot to test the new ABNF, it's 1*2DIGIT, not 2*1DIGIT.

Thanks, fixed.


Bruce will ask about a leading zero in 2DIGIT, is it octal ?
Hm, the RfC 2821 Snum is worse, so maybe that's not critical.

But STD 66 (3968) offers a correct dec-octet, please copy it
for your "qnum".  Or prepare a text about "octal".  Yes, there
is even an RfC about precisely this issue.  Weird or not, it's
how it is.

RFC3968 does look like it is the right RFC.  Which section?



The -02pre2 XML is valid (Bill's validator).  But my qnum test
apparently killed Bill's ABNF checker, I couldn't test the rest
of the ABNF, sorry, bye, Frank

Ha!


Thanks again for the review, it is really appreciated.

Ok, I just got back from a 3hr bike ride, and I'm tired and ready to
head for bed.


-wayne