On 04/12/2016 04:24 PM, Petr Cech wrote:
On 04/12/2016 01:27 PM, Pavel Březina wrote:
> On 04/06/2016 04:55 PM, Petr Cech wrote:
>> On 03/31/2016 03:29 PM, Pavel Březina wrote:
>>> On 03/29/2016 03:17 PM, Petr Cech wrote:
>>
>> Hi,
>>
>> I am going to take three days of PTO. There is my work. Tests don't leak
>> and they really test :-) It is possible to use compact form of them,
>> maybe more help functions... But I hope that main idea is visible.
>>
>> Pavel, you can see TODO comment on line 666. It is a bug. I think that
>> only one rule pass but they all failed. If you see the reason please
>> write me.
>
> Z in the end of the format means UTC+0 timezone and localtime returns tm
> relative to timezone on your machine. Thus you pass UTC+1 but evaluate
> it as UTC+0 and you end up in future time. I think you should be fine if
> you use gmtime instead.
Hi,
fixed patch set attached.
I changed schema from Z to +-0000 (by real localtime). It is still
little wired. So I changed period from 1000 to 10000. It is hack, I know.
I will try to find out if there is bug in sysdb_sudo_convert_time().
And I changed order of includes in Makefile, but I didn't test it on
debian. This was discuss offline.
Regards
--
Petr^4 Čech
Hi,
I investigated the time issue with Pavel -- it is probably known DST
issue [1].
[1]
https://fedorahosted.org/sssd/ticket/2316#comment:1
Patch uses work out with big time shift.
Regards
--
Petr^4 Čech