[krbdev.mit.edu #7532] still not ready for kvnos over 255

[Greg Hudson via RT]

I created a project page describing key version limitations in more 
detail:

http://k5wiki.kerberos.org/wiki/Projects/Larger_key_versions

In addition to the kadmin concern, there are also 16-bit limitations on 
the KDC side.  The proposed changes could risk making our behavior 
worse at 16-bit wraparound than it is currently.  Perhaps this isn't 
worth worrying about; if you rotate a key once per day, you won't hit 
version 32767 until almost 90 years have elapsed.  Regardless, some 
possible approaches are detailed there.