[mail-vet-discuss] abnf tweak

[Tony Hansen]

Dave Crocker wrote:
>>       header = "Authentication-Results:" [CFWS] authserv-id
>>                [CFWS [version]]
>>                *( ";" methodspec
>> 			*( CFWS reasonspec )
>> 			*( CFWS propspec 
> ) [CFWS] ) CRLF
>>       reasonspec = "reason=" dkim-quoted-printable
> 
> 
> A bit of trivial:
> 
> is there a reason for making the optional parameters order-dependent, given that 
> they are self-labeing (and even can occur multiple times)?
>
> If not, I suggest:
> 
>  >                *( ";" methodspec
>  > 			*( CFWS (reasonspec / propspec ) [CFWS] ) CRLF

+1

> I should also ask:  is it really ok to have a methodspec and no reasonspec or 
> propspec?  That's what the current abnf allows, but I'm not sure I understand 
> the semantics.

For many failure cases, there's no discernible property to report on.

	dkim=hardfail reason=d=_and_i=_parameters_missing
	dkim=none

My developers asked me an almost identical question just yesterday. ;-)

	Tony Hansen
	tony at att.com