Taking a look at this, I wonder a bit if we are not over-using 403.  At the
moment we use 403 and a reason for pre-auth failure, ToS failure, and now
for algorithm signature mismatch.
That's a lot of different bike sheds covered with the same paint.  Maybe
that uniformity is good, but , especially in this last case, I think a
regular 400 + reason code is a closer fit (it's not really "forbidden" it's
"fix your syntax").

Is there a strong reason to see this as "forbidden" that I'm not seeing?

(As an individual, in case that's not clear)

Ted

On Tue, Nov 29, 2016 at 1:00 PM, Richard Barnes <[email protected]> wrote:

> As I was chatting with someone about ACME today, I noticed a minor
> problem: The choice of JWS signing algorithm is made by the client, but it
> has to choose an algorithm that the server supports.  I just posted a PR
> that enables the server to refuse a JWS because it doesn't like the
> signature algorithm, and indicate in the error what algorithms it would
> prefer.
>
> https://github.com/ietf-wg-acme/acme/pull/218
>
> _______________________________________________
> 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