On 10/17/2016 07:06 PM, Lukas Slebodnik wrote:
On (14/10/16 13:54), lslebodn wrote:
> URL:
https://github.com/SSSD/sssd/pull/13
> Title: #13: MEMBEROF: Don't resolve members if they are removed
>
> lslebodn commented:
> """
> On (14/10/16 04:48), celestian wrote:
>> I did manual testing with reproducer above. And I ran chmake (it is without
intg., isn't it).
>> Now I check ldap patch with intg.
>>
> Then the question is why manual testing is different than newly added
> integration tests.
>
> BTW It is possible that patch in memberof plugin can safe some unnecessary
> ldb operations and can be considered as perfomance enhancement.
> But it's impossible to say that without proper integration test.
>
Bump,
I am still expecting answer to the comment even though that some
patches were pushed.
LS
Hi Lukas,
the intg. tests in points:
* Adding one user to the group
* let have group with two users -- removing users one by one
* let have user_1 in group_1; user_2 in group_2; group_1 and group_2 in
group_3 -- removing groups (a and 2) one by one from group_3
Manual testing:
# prepare
ipa user-add --first=Adam --last=Adam --email=adam(a)persei.cz adam
ipa group-add group_1
ipa group-add-member --users=adam group_1
ipa group-add group_2
# reproducer
systemctl daemon-reload
sudo su -c "truncate -s0 /var/log/sssd/*.log"
sudo su -c "rm -f /var/lib/sss/db/*"
sudo su -c "rm -f /var/lib/sss/mc/*"
sudo systemctl restart sssd.service
ipa group-add-member --groups=group_1 group_2
sss_cache -UG
sudo su -c "truncate -s0 /var/log/sssd/*.log"
getent group group_2
ipa group-remove-member --groups=group_1 group_2
sss_cache -UG
sudo su -c "truncate -s0 /var/log/sssd/*.log"
getent group group_2
# clean
ipa group-del group_2
ipa group_del group_1
ipa user-del adam
In my opinion intg. tests and manual testing cover the same case.
I wrote 'memberof' patch when I have been working on similiar group
tickets. I assume I confused ticket for `memberof` patch.
Regards
--
Petr^4 Čech