Bug#390696: Authen::Krb5 and get_in_tkt

[Jeff Horwitz]

Thanks Russ. I'll take a look, and if everything looks good I'll make a 
release in the next few days.

-jeff

On Tue, 22 May 2007, Russ Allbery wrote:

> Jeff Horwitz <jeff at smashing.org> writes:
>
>> Hi Russ--
>
>> Once I stopped using Krb5 at the day job, I haven't kept up much with
>> interface changes.  If you can send a patch, I will be more than happy
>> to apply it and make another release.
>
> I finally got around to working on this.  This patch:
>
> * Adds get_init_creds_password and get_init_creds_keytab interfaces that
>   follow the upstream interface and return an Authen::Krb5::Creds object
>   rather than storing the new credentials directly into a ticket cache.
>
> * Adds a destructor for Authen::Krb5::Creds objects now that we're
>   returning them to the user.
>
> * Adds a store_cred() method to Authen::Krb5::Ccache so that the caller
>   can do something useful with the returned credentials.
>
> * Reimplements get_in_tkt_with_password and get_in_tkt_with_keytab in
>   terms of get_init_creds_{password,keytab}, krb5_cc_initialize, and
>   krb5_cc_store_cred.
>
> * Updates the documentation accordingly.
>
> Tested against MIT Kerberos 1.6.1, but these interfaces have been
> consistent for quite some time and shouldn't change with older versions of
> Kerberos.
>
> Let me know if you have any concerns or issues with the patch; I'm happy
> to adjust it as desired.
>
> -- 
> Russ Allbery (rra at debian.org)               <http://www.eyrie.org/~eagle/>
>
>