So to be blunt: You're saying that some of what we want could be achieved
by hacks within the existing model, rather than getting all of what we want
by changing the model :)

I really think it's cleaner at this point to change the model.  There's not
that much deployed code yet, so we should go ahead and get things right.

--Richard

On Thu, Jul 7, 2016 at 6:21 PM, Brad Warren <[email protected]> wrote:

> I think there's a possibility we could implement a lot of the desired
> functionality without preconditions or changing the reg-authz-cert flow.
> The main benefits initially mentioned for preconditions were payments,
> wildcards, and CA issuance flows.
>
> To implement payments, it seems like we could use a generic out of band
> challenge type. I agree that this needs to be more fleshed out.
>
> For wildcards, we could build off of Ron's suggestion to allow clients
> to include a wildcard domain in the authorization request. If we did
> this, however, we'd probably want to modify the HTTP/DNS challenges to
> allow the server to specify the exact domain where the client has to
> complete the challenge.
>
> For example, if the client requests an authorization for *.example.org,
> the ACME server could say that the client needs to complete HTTP
> challenges for a.example.org, b.example.org, and c.example.org.
>
> Changes like those described above would still allow CAs to differ based
> on what the client needs to do to be authorized for an identifier. The
> big downside to this approach is it only allows that flexibility based
> on the domain the client requested. If a CA wants to have different
> preconditions based on values in the CSR other than the domains, this
> won't work.
>
> _______________________________________________
> 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