Debian package comparison

[Russ Allbery]

I'm preparing the next version of the Debian Kerberos packages and went
through all of the patches in the Debian package to make sure that nothing
was missing from the Kerberos RT.

The Debian packages that I'm working on currently contain the patches from
all of the following tickets:

    Kerberos RT #1988
    Kerberos RT #2577
    Kerberos RT #2648
    Kerberos RT #3010
    Kerberos RT #3011
    Kerberos RT #3013
    Kerberos RT #3014
    Kerberos RT #3015
    Kerberos RT #3030
    Kerberos RT #3086
    Kerberos RT #3265
    Kerberos RT #3266

Some of those are new from tonight, but most have been in there for a
while.  All of them have been tested, and all of them should be suitable
for a stable release (although you may not like how the Hurd portability
is handled).

In addition, the Debian package also has the following patches:

bsd-portability

    Patch for GNU/kFreeBSD portability.  This one is brand new; I think
    it should live in the Debian version for at least a version to make
    sure that it really is sufficient.

debian-default-enctype

    Changes the default KDC enctype to DES from triple DES.  I don't
    know/remember the history of this one.

debian-no-rpath

    Disables rpath in binaries and libraries to fit Debian policy.

kprop-service-name

    Patch from Sam from 2001 that makes kpropd and kprop default to a
    hard-coded kprop port if the service wasn't found.  I didn't find this
    one in the Kerberos RT, but wasn't sure if Sam had some reason for not
    already having this one submitted as a bug.  (Or maybe I just missed
    it.)

krb4-bind-des425

    This changes the linkage of libkrb4 to bind tightly (-Wl,-Bsymbolic)
    to libdes425 so that OpenSSL DES libraries won't interpose.  My guess
    is that this is no longer necessary now that everything uses symbol
    versioning, but I'm not positive.

krlogin-alpha

    #undefs TIOCGLTC at the top of krlogin.  I'm not sure of the history
    of this or if it's the right fix in general, rather than just on Linux
    systems.

telnet-help

    Another patch applied by Sam in 2001 that I wasn't sure the status
    of.  It changes the semantics of "telnet help" so that it attempts to
    connect to a host named "help" rather than prints an incorrect usage
    message.

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