-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 08/27/2013 01:31 PM, Simo Sorce wrote:
On Mon, 2013-08-26 at 18:00 -0400, Simo Sorce wrote:
> These patches make use of the krb5.conf for setting the default
> ccache template instead of relying on a pre-configured default
> set a package build time.
>
> The build time parameter is still used as a fallback default if
> krb5.conf and libkrb5 profile library returns no value.
>
> Tested on my machine and works just fine (requires restart of
> sssd if you change krb5.conf).
At Jakub's request on IRC, attached revised patches that also
change the sssd-krb man page and fix the contrib/ patch used by
make rpms.
Patch 0001: Ack
Patch 0002: Nack
As discussed on IRC, please split out the patch that removes the
contrib patch into a separate patch in the set and modify it so that
it reverts to using the built-in defaults (which means updating the
sssd.spec.in as well).
In sss_get_system_ccname_template(), you fail to check the result of
the talloc_strdup(). Also, by convention you should have an explicit
ret = EOK in the success case just before the done: label.
Wouldn't it make more sense to return ENOENT than a NULL ccname? from
sss_get_system_ccname_template()? It's not critical, but we tend to
use that approach in the rest of the code. It would be more clear in
the code what you're doing here too. I had to read it again after
reading the consuming function to understand it properly.
Otherwise, it looks good to me.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.14 (GNU/Linux)
Comment: Using GnuPG with Thunderbird -
http://www.enigmail.net/
iEYEARECAAYFAlIc6lYACgkQeiVVYja6o6OxrwCfaLF2ADX0GfpAokP1Awi4qyIh
fMwAoKDkOhGaCJOeQfUDJ4PMlumVzeAv
=p4PX
-----END PGP SIGNATURE-----