Date: Sun, 31 Dec 1995 14:49:17 -0500
Message-Id: <9512311949.AA22686@portnoy.MIT.EDU>
To: lemson@uiuc.edu (David Lemson)
Cc: krb5-bugs@MIT.EDU, roma@uiuc.edu (Jon Roma), bugs@MIT.EDU
In-Reply-To: "[1763] in Kerberos-V5-bugs"
Subject: Re: KRB5b4.3 & Dec 31
From: John Hawkinson <jhawk@MIT.EDU>
Content-Length: 824

> So it looks like either something is overflowing when it comes to
> 1996 or there is a timebomb somewhere in the code.
>  
> Does anyone have any pointers?  We'll be looking through the code.  :-)

I haven't actually looked at any of the code, but the code-cut Athena
is running started to see this problem (fortunately we don't depend on 
the krb5 stuff for anything).

I _did_ notice that using a kerberos-current or kerberos-b5 kinit
works fine against the same server that fails, so this at least
suggest its exclusively a client problem (harder to fix!).

Beta5 does not seem to have this problem (not surprisingly, given the
above paragraph), and the beta3+SNLsecurid server I have handy also
doesn't exhibit this problem.

This isn't much, but perhaps it'll give you something to against.

--jhawk
  John Hawkinson