Re: [spfbis] Question regarding RFC 7208

["John Levine" <[e-mail]>]

In article <[e-mail]> you write:
>An ABNF like the following would better follow practice in implementations.
>
>    header-field     = "Received-SPF:" [CFWS] result [ FWS comment ]
>                          [ FWS key-value-list ] [FWS] CRLF

While we can have a theological argument about whether the problem is
that the spec doesn't describe the protocol or the programmers didn't follow
the spec, I agree that this ABNF seems more likely to match what people would
consider to be reasonable syntax.


>Is a header field parser allowed to be so robust as to use the latter ABNF to parse the Received-SPF header field?  Is a correction to use that ABNF within the scope of an erratum to RFC 7208?

We are not the protocol police.  You can do whatever you want.
Adjusting your code to handle common benign sender mistakes is often a
good idea.

>(In addition, many cases I've seen include an IPv6 address in the client-ip parameter, which includes colons, without that address being a "quoted-string"
>or fitting the production "dot-atom".  But this issue can wait until I get answers on the main issue I raised in this message.)

I think that's a bug in the spec, due to forgetting that dot-atom doesn't include colons.

R's,
John