[TLS] ESNI/ECH: minor progress, much githubbery

[Stephen Farrell <[e-mail]>]

Hiya,

Today I read over the diff between the latest ESNI/ECH
version and draft-07. [1] I have the following comments:

1. The volume of discussion on github is a deterrent. (*)
I can't keep up with that and coding at the same time
so (being busy elsewhere) paused my coding work in the hope
that the noise would fade over time. Consider this a noise-
reduction plea to allow time for implementation and
testing.

2. Almost all the changes seems fine but near-trivial.
The move to expand/extract from hashes doesn't seem
to add much. I hope there's some theory-justification
for it. I don't see 3 months of significant improvement
so please consider that we may have hit diminishing
returns in the mega-discussion.

3. (This isn't new, but no harm repeating:-) I don't
plan to adhere to the MUST send the public_name from
the ECHConfig. That makes sense for a browser but for
a command line tool, or similar, my conclusion is that
overriding can be justified, so I treat that as a
SHOULD. (I allow such an override for the openssl
s_client version I've done and similarly for curl.)

FWIW, I hope to now have time to resume coding. I won't
have time to process the volume of mails generated by
recent github discussion as well so plan to pay attention
to the text, openssl code and the mailing list. I
might or might not notice something significant in
the githubbery, so would be happy if significant changes
(if any) can be sent to the list.

Cheers,
S.

[1]
https://tools.ietf.org/rfcdiff?url1=https://tools.ietf.org/id/draft-ietf-tls-esni.txt&url2=https://tlswg.github.io/draft-ietf-tls-esni/draft-ietf-tls-esni.txt#part-2


(*) I get mail when there are comments on that repo. I
have gotten 784 since June 1 when draft-07 was published.
There were also 557 github emanations related to SVCB
in the same timeframe.