Hiya,

On 23/08/2022 22:51, Brian Dickson wrote:
The differences in interpretation, and the client behavior under one of
those interpretations, are the problem.

I've seen a different client-behaviour issue related to ports
other than 443 and ECH, but I'm unsure if that's a problem
with this spec or simply an implementation issue that'll be
fixed in short order.

Currently chromium and firefox disagree on whether ECH is
setup correctly for one of my test pages [1]. Initially, I
had misinterpreted what to publish in an HTTPS RR for [1] and
neither browser liked that, but even after I fixed that,
(thanks to one of the browser folk educating me), only one
seems to work, and the other still thinks that ECH isn't
properly setup for [1]. Surprisingly though, both work for
another page [2], but I wonder if that may be because ECH is
also enabled for the same name on the default port. [3]

If people are going to take another look at this spec, it may
be useful to also see if the text relating to ports other
than 443 is sufficiently clear - I know I got what to publish
wrong, and the fact that browsers haven't yet landed on the
same interpretation of what's needed for ECH away from port
443, may indicate a bit more clarity would be beneficial.

To be clear: I'd be fine with the RFC being issued and us
figuring out any useful clarifications as experiments with
ECH continue over the next while - my guess is there'll be
more than just this aspect of ECH to document better.

Cheers,
S.

PS: In case you're clicking the links below - both browsers
require a recent(ish) build, use of DoH and turning on a flag
before they attempt ECH so you'll only see the difference if
you've all that setup on your client.

PPS: Some of the relevant folk were vacating recently (me
included) so it could be this just gets fixed in the very
near future, but I figured if there's going to be a window
when some editorial/clarity text improvements might be made
it was worth raising, just in case.

[1] https://draft-13.esni.defo.ie:8413/stats
[2] https://my-own.net:8443/ech-check.php
[3] https://my-own.net/ech-check.php

Attachment: OpenPGP_0x5AB2FAF17B172BEA.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature
Description: OpenPGP digital signature

_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to