[983] in Zephyr_Bugs
Re: Zephyr snapshot
daemon@ATHENA.MIT.EDU (Steve Langasek)
Wed Dec 22 12:28:43 1999
Date: Wed, 22 Dec 1999 11:28:40 -0600 (CST)
From: Steve Langasek <vorlon@dodds.net>
To: Greg Hudson <ghudson@MIT.EDU>
Cc: zephyr-bugs@MIT.EDU
In-Reply-To: <199912221625.LAA10051@small-gods.mit.edu>
Message-Id: <Pine.LNX.4.10.9912221121550.24220-100000@norad.dodds.net>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Wed, 22 Dec 1999, Greg Hudson wrote:
> I'm afraid your patches aren't correct.
> The change of "mv" to "cp" for zephyr_err.h is a bit perplexing; I
> don't see what it has to do with kth Kerberos.
kth-krb4 ships includes libcom_err and compile_et, which behave
differently than the MIT version. The result is that moving zephyr_err.h,
upon which zephyr_err.c depends, to h/zephyr/ causes the build to fail on
zephyr_err.c.
> The checks for krb_set_key(), decomp_tkt(), and krb_rd_req() are
> misguided. These functions appear in MIT Kerberos as well, and are
> modified in Zephyr. Which is really very ugly (it's not my fault,
> certainly), and may very well break interoperability with kth
> Kerberos, but your patch doesn't have the right effect.
Ah. In that case, some header work will have to be done, because there are
some incompatible prototypes being included indirectly from kth-krb4.
I'll look into what will be needed to get this working.
-Steve Langasek
postmodern programmer
