On Wed, May 12, 2010 at 03:58:23PM -0700, Love Hörnquist Åstrand wrote:
> 
> 12 maj 2010 kl. 15:37 skrev Peter Saint-Andre:
> 
> >> So I'm not sure right now what to say about that. I suspect we can still
> >> stipulate that the only RDN having attr type of CN that we'll pay
> >> attention to is the one at the far end of the RDN sequence comprising
> >> the DN.
> > 
> > We can stipulate that, but is it realistic?
> 
> Yes, since that's what RFC 2818 said.

RFC 2818 says that you use the most specific CN RDN.  Not the most 
specific RDN, iff it is a CN.  That is completely different.

The stats I ran said that maybe 10/20% of existing certs place the CN 
somewhere other than the most specific field.  Who out there is going to 
change their software to implement a CN matching algorithm which breaks 
interop with 10/20% of SSL servers?

Regards, Joe
_______________________________________________
certid mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/certid

Reply via email to