On 3/6/25 09:18, Daniel K. wrote:
> ; domain-name UNDEFINED
>
> It's clear what domain-name is supposed to be, but it being
> syntactically undefined is problematic.
>
> I suggest we add this to resolve it:
>
> domain-name = dot-atom-text ; from RFC5322
I have been digging into this a bit more and I found that in RFC 7489,
we referenced RFC 6376 (DKIM) as the authority for "domain-name". This
reference was lost along the way, somehow.
Two other places in the draft we refer to "domain" from RFC 5322. In
other words, we are using two competing definitions for a domain name.
I suggest we replace "domain" from RFC 5322, as it is very liberal in
what it accepts (basically "atext"), with the more conservative
"domain-name" from RFC 6376 for the affected rules involved in the
construction of "filename".
filename = receiver "!" policy-domain "!" begin-timestamp
"!" end-timestamp [ "!" unique-id ] "." extension
- receiver = domain
- ; imported from [@!RFC5322]
+ receiver = domain-name
+ ; imported from RFC 6376
- policy-domain = domain
+ policy-domain = domain-name
After this, and some other fixes I will submit a pull request for, the
ABNF tool have the following to say:
; dot-atom-text UNDEFINED
; DIGIT UNDEFINED
; ALPHA UNDEFINED
; FWS UNDEFINED
; filename defined but not used
; dmarc-subject defined but not used
All the UNDEFINED tokens are resolved via a comment in the ABNF or the
surrounding prose.
The defined but not used ones, are referred to in the prose.
Daniel K.
--
last-call mailing list -- last-call@xxxxxxxx
To unsubscribe send an email to last-call-leave@xxxxxxxx
