Implementing internationalization

[Sam Hartman]

Folks, there is a discussion of combining draft-ietf-krb-wg-preauth-fw
and i18n and new ASN.1 in krb-wg.

I want to give that working group an idea of what it would take for
MIT to be able to reasonably implement that proposal.


The big problem for us would be if it caused us to change public
structs.


Presumably we'd have to decide that strings were either locale
specific or utf8 possibly as a context option.

So, things we could do:

* add  extensibility markers

* Tag strings differently

* Add typed holes that our implementation can ignore.

Things it would be difficult to do:

* Add fields we care about to ap-req or ap-rep

*Add fields we care about to the ticket structure


* Add fields we care about to the encrypted ticket structure

* Add fields to the kdc-req or kdc-rep.

* Add fields to krb_cred or its encrypted part.

Basically we can add fields to those provided that our implementation
can completely ignore them.

Does this seem right to people.

--Sam