[ietf-dkim] Errata for RFC4871

[Tony Hansen]

Pasi.Eronen at nokia.com wrote:
> As was noted in the DKIM meeting, there are currently 12 reported
> errata for RFC 4871:
> 
> http://www.rfc-editor.org/errata_search.php?rfc=4871
> 
> Couple of questions:
> 
> Errata ID 1383: I found some old emails (April 2008) suggesting that
> the proposed change might be actually wrong (ABNF doesn't allow 
> multiple wildcards in the string). If so, a new errata should be 
> submitted.

Pasi is correct here: the errata is wrong with respect to multiple
wildcards.

However, it would still be useful to add an example of foo*bar.

This was prompted by someone at the interop noting that some people had
not coded for the example of a * in the middle.

> Errata ID 1378: This looks like a a technical change to text that's
> relatively unambiguous ("REQUIRED").  I'm wondering what existing
> implementations do here? If everyone includes the "a=" parameter, it
> could be better to clarify Section 3.3 instead (removing the sentence
> "..if no algorithm is specified").  And do existing implementations
> assume the "a=" parameter is present?  (I don't have much information
> either way -- but I guess the WG does.)

This could go either way. As the errata says, section 3.3 of 4871 says
that there's a default if a= is not specified, and section 3.5 says it's
required.

As the Errata says: we need to pick one. Either 1) remove the sentence
from section 3.3, or 2) change "REQUIRED" to "OPTIONAL" in section 3.5.

I don't care which.

> I'd like the WG to review all the errata. Once the WG chairs send
> a list of errata IDs where there's rough WG consensus that the proposed 
> changes are correct, and appropriate to do via the errata process, I'll 
> send approval to the RFC editor.

+1

	Tony Hansen
	tony at att.com

> Best regards,
> Pasi
> 
> _______________________________________________
> NOTE WELL: This list operates according to 
> http://mipassoc.org/dkim/ietf-list-rules.html