[TLS] Re: 2nd Working Group Last Call for The SSLKEYLOGFILE Format for TLS

[Stephen Farrell <[e-mail]>]

I think we should abandon this effort and not publish.

When initially proposed this was supposed to be documenting a
deployed reality. I could just about hold my nose with that,
but adding in ECH (for which key exfiltration is not currently
deployed, nor seemingly needed) and the IANA considerations make
this much more unacceptable.

While I expected to be in the rough for the initial proposal I
do hope others find this as unacceptable as I do and we don't
publish.

Including ECH here is IMO not necessary - there is no history
of needing that for debugging purposes so adding it based on
speculation is IMO wrong and a fatal error.

This is now extensible (via IANA specification required) which
is a problem as that means anyone can likely define new ways
to exfiltrate secrets from TLS implementations without any WG
overview. That IMO is another fatal error.

Less importantly, but still substantively, the draft does not,
but should, recommend that this feature only be available via
conditional compilation (where available) and not be part of
any standard library or other release. If publishing, we should
be aiming for the strongest possible implementation guidance
in that respect.

Thanks,
S.