getprinc problems in 1.8b1
mdw at umich.edu
Wed Feb 24 14:20:43 EST 2010
> Date: Wed, 24 Feb 2010 13:44:41 EST
> To: Marcus Watts <mdw at umich.edu>
> cc: "krbdev at mit.edu" <krbdev at mit.edu>
> From: Greg Hudson <ghudson at MIT.EDU>
> Subject: Re: getprinc problems in 1.8b1
> On Wed, 2010-02-24 at 11:55 -0500, Marcus Watts wrote:
> > The second has to do with mkvno. This value is always
> > returned, regardless of whether KADM5_MKVNO is set.
> > This was new behavior in 1.7. (1.6.3 had different
> > possibly broken behavior.)
> Out of curiosity, did this cause a problem?
> Your reasoning and patch seem sound, so I'll commit this and tag for
> pullup to 1.8. (It might conceivably be delayed until 1.8.1, but we
> expect to roll 1.8.1 much sooner after 1.8 than we did 1.7.1 after 1.7.)
It didn't cause a problem per se - it was just "weird" and unexpected.
Of course, I don't (yet) have any reason to care about mkvno. If I
had a procedure that incrementally changed mkvno, I might then
become very interested in a way to see what mkvno was.
I think the behavior of this must have been a concern for the maintainers
of Authen::Krb5::Admin . They do support KADM5_MKVNO, including a method
call "mkvno", but they don't document it. My guess is they couldn't figure
out what was happening, and decided it was simpler to just not document it.
More information about the krbdev