Change in K4 ticket cache format

[Russ Allbery]

Sam Hartman <hartmans at mit.edu> writes:
>>>>>> "Russ" == Russ Allbery <rra at stanford.edu> writes:

>     Russ> Well, in 1.2 everything used KRB4_32 uniformly.  It was only
>     Russ> in 1.3 that the read and write functions switched back to a
>     Russ> long.  (Maybe for compatibility with old krb4 libraries?)

> Can you please confirm you are not dealing with Stanford local
> patches?

> I think that's what we determined the last time we looked at this.

Well, I *thought* I had a virgin tarball, since that would have been the
whole point of having both a 1.2.8su and a 1.2.8.  However, the "virgin"
1.2.8, er, wasn't.  *sigh*  Sorry about that; I was operating on false
assumptions.

I just downloaded a really virgin 1.2.8 tarball from web.mit.edu, and
indeed, you're both entirely correct and long was always used in the
ticket read/write routines.  We must have "fixed" this years back, before
I was working on Kerberos code, and never got those patches accepted back.
Sorry about the wild goose chase; I'll make sure the institutional memory
hangs on to the fact *this* time.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>