Microsoft Referral Code for Clients

Kevin Coffman kwc at
Thu Jul 10 10:37:13 EDT 2003

I remember that one concern was that the original patch only allowed 
referrals to a single other realm.  We had a need to do referrals to 
both a test forest root or our production forest root.  We changed the 
patch to attempt to determine where the referral should go by looking 
at the domain name of the requested service.  However, there are cases 
where Windows sends a short name without domain information.  So we 
punted and changed the code again to have a default realm where clients 
are referred if the domain cannot be determined otherwise.

I have not heard of any problems with this setup here.

I'm not sure if there were more concerns with the original patch, 
besides it not being part of the standard at the time.


> When you say concerns, what do you mean?
> -----Original Message-----
> From: Kevin Coffman [mailto:kwc at] 
> Sent: Wednesday, July 09, 2003 8:59 PM
> To: 'Sam Hartman'; 'Wachdorf, Daniel R'
> Cc: krbdev at
> Subject: RE: Microsoft Referral Code for Clients
> > 
> > 6) I'm not sure we would accept client side patches without
> >    server-side patches.  Umich has  not submitted their server-side
> >    patches for  inclusion and we know of no plans on their part to do
> >    so.
> Our patch is a modification of the original patch that, AFAIK, originated at
> Microsoft but came to us from MIT.  It was my understanding that there were
> (valid) concerns with the original patch.  I'm not positive all the concerns
> have been addressed by our changes.  We're *more* than willing to have the
> patch (or at least its functionality) in the MIT distribution. Let me know
> if I should submit the patch via RT.

More information about the krbdev mailing list