Change in K4 ticket cache format

[Alejandro R. Sedeno]

On Thu, 2005-05-19 at 14:53 -0400, Jeffrey Hutzelman wrote:
> 
> On Thursday, May 19, 2005 01:27:58 PM -0400 Sam Hartman <hartmans at mit.edu> 
> wrote:
> 
> > I think an approach like this would be worth accepting probably even
> > if it only works for little endian systems.  I don't think we at MIT
> > have time to adapt this patch.
> I can't remember if I said this already, but...
> While I haven't found a way to design an "alignment record" that will work 
> on big-endian systems, it should still be straightforward to make the 
> ticket file code able to _read_ files written in either format (or a 
> mixture).

After being bitten by this and talking to Sam, I'm planning to port
Jeff's patch to MIT Kerberos.

My plan is to have us write out a 32 bit issue_date, and the "alignment
record" as suggested by Jeff's patch. The record will be ignored, but
will not be optional just yet. I understand that may have to change
before the patch is accepted, but I don't want to deal with it just yet.

The first half of this change (32 bit issue_dates) is done and at 

http://web.mit.edu/asedeno/www/krb4-tkt-fix.diff

I expect the alignment record will be done later tonight, though I
wouldn't mind any feedback in the mean time.

-Alejandro