On Fri, Nov 14, 2014 at 10:20:51AM +0100, Pavel Reichl wrote:
On 11/13/2014 07:25 PM, Lukas Slebodnik wrote:
>On (13/11/14 19:06), Pavel Reichl wrote:
>>On 11/13/2014 06:57 PM, Jakub Hrozek wrote:
>>>On Thu, Nov 06, 2014 at 12:26:41PM +0100, Sumit Bose wrote:
>>>>this patch adds a new sysdb call which is needed to support overrides
>>>>for IPA users and groups.
>>>ACK except for some weird indentation at the end of the unit test. I can
>>>fix that before pushing.
>>>>Pavel, I tried to make the new call match the changes you are currently
>>>>working on. Can you check if it is really the case?
>>>Looks OK, but I'll let Pavel answer that question..
>>LGTM, result->count == 0 implies ENOENT to be returned and output result is
>>undefined (should not be accessed if return value is not EOK).
>I thought you introduce SSS_LDB_SEARCH to simplify these things.
I did but I don't demand its usage. :-)
I understand objections against using macro and I think this code looks OK
There's a lot of code that behaves according to the policy I described above
and does not use the macro.
I don't think there is going to be an afford to use it everywhere, so I
don't care if macro is used or not the policy is what I care about.
I'm not sure if SSS_LDB_SEARCH would be more readable here since the error
checking block after the ldb_search() also checks for more than one