[PATCH] LDAP: handle SID requests if noexist_delete is set

[Sumit Bose]

Hi,

I found this while testing the cifs-utils plugin I opened
https://fedorahosted.org/sssd/ticket/2116 to make sure all affected
versions will be fixed.

bye,
Sumit

[Jakub Hrozek]

On Wed, Oct 09, 2013 at 10:44:49PM +0200, Sumit Bose wrote:
> Hi,
>
> I found this while testing the cifs-utils plugin I opened
> https://fedorahosted.org/sssd/ticket/2116 to make sure all affected
> versions will be fixed.
>
> bye,
> Sumit

This patch fixes the problem, thank you. I tested with
pysss_nss_idmap.getnamebysid against an IPA server. But I found another
potential issue -- if the entry is not found, we don't remove it in the
sdap_get_user_and_group() request. This means the entry can stay in the
cache potentially forever even after it's removed from the server, if it's
only looked up by SID.

So ACK to your patch but I'll send another one to address the deletion.

[Jakub Hrozek]

On Thu, Oct 10, 2013 at 07:37:58PM +0200, Jakub Hrozek wrote:
> On Wed, Oct 09, 2013 at 10:44:49PM +0200, Sumit Bose wrote:
> > Hi,
> >
> > I found this while testing the cifs-utils plugin I opened
> > https://fedorahosted.org/sssd/ticket/2116 to make sure all affected
> > versions will be fixed.
> >
> > bye,
> > Sumit
>
> This patch fixes the problem, thank you. I tested with
> pysss_nss_idmap.getnamebysid against an IPA server. But I found another
> potential issue -- if the entry is not found, we don't remove it in the
> sdap_get_user_and_group() request. This means the entry can stay in the
> cache potentially forever even after it's removed from the server, if it's
> only looked up by SID.
>
> So ACK to your patch but I'll send another one to address the deletion.

Pushed to master and sssd-1-11