RE: I-D ACTION:draft-ietf-secsh-transport-18.txt

[pgut001%cs.auckland.ac.nz@localhost (Peter Gutmann)]

Scott Rankin <scottra%wrq.com@localhost> writes:

>x509v3-sign-rsa and x509v3-sign-dss are listed as "defined" formats. That
>said, I have been unable to find where these are defined (and there is no
>citation of this definition in this section).

The last time this came up I dug through the entire archives and posted a
summary of all the posts I could find on the topic (it'll be in the archives
somewhere, grep for my From: address).  The conclusion was that no-one could
agree on how this should work, and no-one was interested enough to sit down
and sort it out.  Unfortunately those two ID strings are still present in the
spec although they're not useful for anything, I think it'd be best to either
remove them entirely (make them accessible via a vendor-specific string if
anyone does actually want to use them), or just include a note to say that
these values exist for historical purposes but aren't used for anything, sort
of like the "entry" keyword in K&R C.

Peter.