Re: [Doh] draft-ietf-doh-dns-over-https-04 feedback

Thanks Ray,

imo pretty much this is what SHOULD already signals - I would phrase it as:
that's the expected behavior unless you have a really good reason you
cannot (Dave points out one really good reason - conflicting non-successful
states). Its also plausible an HTTP gateway would not have enough
information to determine more than "failure" and would want to fallback on
400 which would be less than optimal but ok. This is sort of a general
problem and I don't think the draft is overall advantaged by doing this
case by case.

on the other hand we did add some similar text around success/non-success
and how dns and http interact. That's also not strictly necessary but I
think it has a broad enough impact that it is worth doing. I would hope to
avoid doing that with a laundry list of codes (my approach is consistent
with BCP56bis )

On Thu, Mar 22, 2018 at 10:47 AM, Ray Bellis <ray@bellis.me.uk> wrote:

> On 22/03/2018 10:32, Dave Lawrence wrote:
>
> > Because it might have good reasons for replying with some other code,
> > if multiple issues exist with the request?
>
> OK, so you need to allow for HTTP protocol layer errors such as 401, etc
> which happen earlier in the stream's processing.
>
> But if it gets that far, is there any other possible response than 415?
>
> I'm not sure how to phrase it, but I think it MUST be 415 *unless* some
> other error was *already* detected (at which point I'd expect request
> processing to have been aborted anyway).
>
> Ray
>
> _______________________________________________
> Doh mailing list
> Doh@ietf.org
> https://www.ietf.org/mailman/listinfo/doh
>