Agreed. The removal greatly simplifies the protocol. As you noted, the
addition of the "application requirements" achieves the same intended
result.

On Wed, Aug 17, 2016 at 12:41 PM, Richard Barnes <[email protected]> wrote:

> SGTM.  I never like "combinations" much anyway :)  I put one editorial
> comment in the PR.
>
>
> On Wed, Aug 17, 2016 at 2:22 PM, Jacob Hoffman-Andrews <[email protected]>
> wrote:
>
>> https://github.com/ietf-wg-acme/acme/pull/171
>>
>> This is a fairly complicated part of the protocol, and not used in
>> practice. For instance, in Let's Encrypt's implementation, there are
>> always three challenges, any one of which may be fulfilled by the client.
>>
>> After this change, all challenges are considered to be combined with an
>> "OR." That is, any challenge within an authorization may be completed to
>> make the authorization valid.
>>
>> Authorizations within the new-application object are considered to be
>> combined with an "AND." That is, all of them must become valid before
>> the certificate will be issued. The combination of the two means that we
>> have similar expressiveness as before, even without the combinations
>> array.
>>
>> Thoughts?
>>
>> _______________________________________________
>> Acme mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/acme
>>
>
>
> _______________________________________________
> Acme mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/acme
>
>
_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to