The suggested changes look good... just a few comments in-line.

On 1/7/26 7:36 PM, Tobias Looker wrote:

>> 1155 The Relying Party SHOULD send the following Accept HTTP Header to
>> 1156 indicate the requested response type unless the Content-Type of
>> 1157 Status List Tokens in the respective ecosystem is known or the
>> 1158 Relying Party supports both formats:
>> 
>> 1160 * "application/statuslist+jwt" for Status List Token in JWT format
>> 
>> 1162 * "application/statuslist+cwt" for Status List Token in CWT format
>> 
>> 1164 If the Relying Party does not send an Accept Header, the response
>> 1165 type is assumed to be known implicitly or out-of-band.
>> 
>> Is this using something other than normal HTTP content negotiation? If not, I
>> think it is better to identify the media types for the status list formats 
>> and
>> defer to how HTTP does content negotiation.
> 
> Our intention was to use standard HTTP content negotiation. If that is not 
> clear from the text, we can change the section to just define the media types 
> and point to rfc9110?

I think that would be better. Thanks.

>> Redirection
>> 1557 HTTP clients that follow 3xx (Redirection) class of status codes
>> 1558 SHOULD be aware of the possible dangers of redirects, such as
>> 1559 infinite redirection loops, since they can be used for denial of
>> 1560 service attacks on clients. A client SHOULD detect and intervene in
>> 1561 infinite redirections. Clients SHOULD apply the guidance for
>> 1562 redirects given in Section 15.4 of [RFC9110].
>> 
>> Why aren't these MUST? Is there a reasonable scenario in which a client is
>> advised to be unaware of infinite redirection loops, etc...?
> 
> In our opinion, the first should can be non-normative, but the decision for a 
> client to implement detection etc. also heavily depends on trust assumptions 
> etc. In cases where the client fully trusts the Status List Issuer, it might 
> not be necessary to implement such measures. We were also not entirely 
> certain how easy this is to implement in practice (or well supported in 
> libraries).

My experience is that some libraries do detect it and some don't. However, even 
in a high-trust environment it is advisable to detect redirection loops as both 
mistakes and compromises happen.
I won't belabor this issue if you feel strongly.

-andy, ART AD

_______________________________________________
OAuth mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to