John C Klensin wrote on the SMTP list:
I would encourage checking the "Change Log" appendix and/or
doing a diff against -08.
No problems in the diff. Checking -09 against the second LC
comment <http://article.gmane.org/gmane.ietf.general/29277>:
(1)
| In section 4.1.3 <dcontent> is undefined
[...proposed fix...]
RFC 2822 offers <dcontent> with <NO-WS-CTL>, the normative
reference 2822upd-06 doesn't have a <dcontent>. You're not
planning to fix this with an erratum, or are you ?
(2)
| <Standardized-tag> got no valid syntax
[...proposed fix...]
(3)
| For an <IPv6-addr> importing or adapting <IPv6address>
| from STD 66 would be clearer: The STD 66 syntax works
| as is, without any additional restrictions in comments.
(4)
| In section 4.2 <text> is undefined
[...proposed fix...]
(5)
| In section 4.2.2 (reply codes by function group) 251 and
| 551 should form one group
(6)
| In section 4.1.2 <Keyword> and <Argument> are apparently
| unused and can be removed. For the collected syntax see
| <http://hmdmhdfmhdjmzdtjmzdtzktdkztdjz.googlepages.com/2821bis.ABNF.1>
(7)
| The ehlo-greet = 1*(%d0-9 / %d11-12 / %d14-127) allows
| NUL and obs-NO-WS-CTL
If that is intentionally the NULs could get their very own
security consideration.
(8)
Somebody asked where the 2821 Interop and Implemementation
report mentioned in the Last Call is. My own tests in 2007
only showed that (1) <NO-WS-CTL> in <dcontent> doesn't work,
I didn't test (7). In 2007 I forwarded Dave's request for
reports to the ASRG, RFCI, and SPF devel mailing lists.
Frank