-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 11/22/2010 04:03 AM, Sumit Bose wrote:
On Fri, Nov 19, 2010 at 02:23:21PM -0500, Stephen Gallagher wrote:
All errnum values passed into this function throughout the code
are PAM error codes, but we were passing them through strerror()
to print them, which is only meaningful for ERRNO error codes.
This patch changes dp_err_to_string() to use pam_strerror().
> Maybe it makes sense to rename it to something like
> dp_pam_err_to_string() ?
Done. New patch attached.
Fixes
https://fedorahosted.org/sssd/ticket/636
Note: the first argument to pam_strerror() requiring a PAM handle is
historical and not necessary. I tested that passing it NULL results in
no unexpected behavior.
> Although I'm not aware of any example I would suggest to add a configure
> test to make sure that there is no PAM implementation which assumes that
> the first argument is a valid PAM handle. But I would agree to add this
> as a new enhancement ticket, because it will be only important if we
> start to test sssd on other platforms than Linux.
I don't think there's actually any way to detect this, as the definition
for this function allows no error code returns (just a valid constant
string or NULL, if the error code can't be mapped to one).
I suppose it's possible that it could segfault on certain platforms, but
it doesn't seem likely.
I'm inclined to ignore this unless we actually find a platform that hits it.
- --
Stephen Gallagher
RHCE 804006346421761
Delivering value year after year.
Red Hat ranks #1 in value among software vendors.
http://www.redhat.com/promo/vendor/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Fedora -
http://enigmail.mozdev.org/
iEYEARECAAYFAkzqkPAACgkQeiVVYja6o6NGagCfVP7CfmiJPaZeiEoWUIprEudv
DHoAniRBtAxlt3j/CzMz+5JySBsOru8h
=4cdt
-----END PGP SIGNATURE-----