Re: [dmarc-ietf] Fwd: [Technical Errata Reported] RFC7489 (7865)

[Alessandro Vesely <[e-mail]>]

On Sun 24/Mar/2024 13:33:22 +0100 Alessandro Vesely wrote:
> On Sat 23/Mar/2024 19:53:39 +0100 John Levine wrote:
>> It appears that Murray S. Kucherawy  <[e-mail]> said:
>>> -=-=-=-=-=-
>>>
>>> This seems like it's probably legitimate.  Does it need to be fixed in the
>>> -bis document?
>>
>> It's already fixed in the current markdown.
>>
>> FYI, the XML pattern is silly.  It forbids harmless stuff like leading zeros 
>> in 01.02.03.04
>> and doesn't allow some exotic but valid IPv6 forms like ::ffff:12.34.56.78.
> 
> 
> How about:
> "(::ffff:)?(([01]?\d?\d|2[0-4]\d|25[0-5])\.){3}([01]?\d?\d|2[0-4]\d|25[0-5])"/>


Testing yielded a correct fix:

   "(::[Ff]{4}:)?(([01]?\d?\d|2[0-4]\d|25[0-5])\.){3}([01]?\d?\d|2[0-4]\d|25[0-5])"/>


I'd agree it might have been better to allow an overly lax pattern, such as "[0-9a-fA-F.:]+".  When we (Tim and I) got to revise the grammar from Freddie's work on docs.google.com, in March 2020, we tried and followed the style of RFC 7489.  That led to the current grammar under the comment <!-- The Internet Protocol Address from which messages were received -->.  Obviously we hadn't tested all cases.  Now you (John) found a case.  Why shouldn't we fix it?  Why would it be too late, since we're in WGLC?  Why would it be the wrong place?

Note: the change doesn't make the grammar stricter than it is.  It makes it slightly more relaxed.

Please take my pull request.


Best
Ale
--