From: Theo Buehler <tb@theobuehler.org>
Subject: Re: rpki-client: collect non-functional CAs
To: Job Snijders <job@openbsd.org>
Cc: tech@openbsd.org
Date: Wed, 12 Mar 2025 12:15:45 +0100

On Wed, Mar 12, 2025 at 10:51:54AM +0000, Job Snijders wrote:
> On Wed, Mar 12, 2025 at 10:02:04AM +0100, Theo Buehler wrote:
> > +	json_do_array("nonfunc_cas");
> > +	RB_FOREACH(nca, nca_tree, ncas) {
> > +		json_do_object("nca", 1);
> > +		json_do_string("location", nca->location);
> > +		json_do_string("ta", taldescs[nca->talid]);
> > +		json_do_string("caRepository", nca->carepo);
> > +		json_do_string("rpkiManifest", nca->mfturi);
> > +		json_do_string("ski", nca->ski);
> 
> I think I'd prefer using pretty_key_id() to format the SKI. (In a
> separate diff we can use pretty printing for the SKI in BGPsec too.)

It seems a bit strange to dump a hex string with colons into the JSON.
Isn't making things easy on the eye the job of whatever software
displays this to a human?

But if you do want to do that, I'd suggest you switch the BGPsec SKI now
(with a clear commit message so I don't forget to call out the change in
the ChangeLog) and I'll match the change in my tree.

> Tested, very cool results. Thanks!

Glad you like it. Claudio wants to take a closer look, so I'll wait for
him.