On Fri, May 20, 2022 at 6:08 AM Dmitry Belyavsky <[email protected]> wrote:

> On Tue, May 17, 2022 at 7:28 PM Murray S. Kucherawy <[email protected]>
> wrote:
>
>> 1) In Section 3, you have:
>>
>> "The validation rules introduced in RFC 6531 are considered to be
>> followed."
>>
>> I don't quite understand this sentence.  Do you mean this?
>>
>> "It is assumed that addresses used with this extension will pass the
>> validation rules introduced in RFC 6531."
>>
>> If not, please clarify.
>>
>
> Yes, we mean this, but we would prefer the following wording:
> "The validation rules introduced in RFC 6531 MUST be followed when
> processing this extension."
>
> Are you OK with this?
>

Yes, thanks, that's clearer.


>
>>
>> 3) For Section 8, I suggest this to get rid of the layered
>> SHOULD/RECOMMENDED:
>>
>> "To reduce the risk of future usability errors, registries SHOULD
>> validate all code points in the domain name of any provided email address
>> according to IDNA2008 [RFC5892]."
>>
>> Then again, usability errors aren't something I would expect to be
>> discussed in a Security Considerations section, so maybe this should be
>> someplace else?
>>
>
> Would you like something like this?
>
> "As email address is often a primary end user contact, invalid email
> address may put the communication with the end user into risk in case when
> such contact is necessary. To reduce the risk of the use of invalid domain
> names in email addresses, registries SHOULD validate the domain name syntax
> in the provided email addresses and validate all code points in the domain
> name according to IDNA2008 [RFC5892]"
>

That's better, but this still doesn't feel like a security matter to me as
worded.  If you want to add a sentence or two about what security threat
exists if this validation isn't done, that makes a better argument.


>
>>
>> 4) You might want to say something explicit about all of the EAI security
>> issues also applying to this work.
>>
>
> We have pretty well described security considerations in RFCs 6530 and
> 6531. I think referring to them is a good option. I don't think we have any
> extra security considerations here.
>

Fine with me.


> If you are OK with the suggested changes, I will publish the updated draft
> version ASAP.
>

Yep, ship it!

-MSK
_______________________________________________
regext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/regext

Reply via email to