Unicode and APIs

[Sam Hartman]

>>>>> "Russ" == Russ Allbery <rra at stanford.edu> writes:

    Russ> Sam Hartman <hartmans at MIT.EDU> writes:
    >> Ken mentioned in another post that he suspected that adding
    >> internationalization support would require significant API
    >> work.

    >> I'd like to challenge that assumption.

    >> How badly will things break if we add a context flag that says
    >> "everything is UTF8" and minimize other API changes?

    Russ> I think you still have to deal with IDN for any place where
    Russ> a hostname is embedded in a Kerberos principal name and
    Russ> where automatic discovery of the appropriate principal name
    Russ> for a service is done.

I agree.  Assuming that we decide that hostnames in kerberos
principals are ACE-encoded then you want to convert anything that
looks unicode (assuming the utf8 flag is set) that comes into sname_to_princ or anything else that accepts a hostname into ACE.

If we decide that principals are not ACE, well, then you want to
accept ACE and UTF8|current locale.