[1323] in Kerberos-V5-bugs
Re: krb5b4pl3: lib/krb425/rd_req.c: "*" instance should use first
daemon@ATHENA.MIT.EDU (Theodore Ts'o)
Fri Mar 31 17:03:15 1995
Date: Fri, 31 Mar 1995 17:02:15 +0500
From: Theodore Ts'o <tytso@MIT.EDU>
To: Marc Horowitz <marc@MIT.EDU>
Cc: Theodore Ts'o <tytso@MIT.EDU>, krb5-bugs@MIT.EDU,
"Jonathan I.
Kamens" <jik@cam.ov.com>
In-Reply-To: [1295]
[1295] daemon@ATHENA.MIT.EDU (Marc Horowitz) Kerberos-V5-bugs 03/24/95 22:17 (13 lines)
Date: Fri, 24 Mar 1995 22:16:57 EST
From: Marc Horowitz <marc@MIT.EDU>
>> (You guys aren't actually using it for anything, are you?)
We use zephyr linked against it, rather than rewriting to the krb5
api. (Of course, we could do this, but frobbing the imake config was
easier.) I don't think you should be removing parts of the release
which are useful just because they can be misused. Who is abusing it?
Who has misused it in the past? Several people, at places from Aragonne
National Labs, Sandia National Labs, and Rutgers, from a quick glance
through my mail files. Basically, they think of it as a replacement
krb4 library, which is most definitely is not, and then they send in
long complicated patches to try to make it do it.
My main reason for removing it is that I don't want to encourage people
to use it as an interface. It's a bad idea, if what you're looking for
is portability and interoperability at the protocol layer. (I try to
pursuade people not to use sendauth/recvauth for the same reason, now
that GSSAPI is available --- especially if they are writing new
applications.)
I could be convinced to put it into an unsupported contrib directory.
- Ted
