URL:
https://github.com/SSSD/sssd/pull/59
Title: #59: ipa_netgroups: Lowercase key to htable
jhrozek commented:
"""
On Thu, Oct 20, 2016 at 02:55:12AM -0700, lslebodn wrote:
On (20/10/16 02:20), Jakub Hrozek wrote:
>On Thu, Oct 20, 2016 at 01:40:15AM -0700, lslebodn wrote:
>> On (20/10/16 01:21), Jakub Hrozek wrote:
>> >before the patch:
>> >```
>> >[jhrozek@client] sssd $ [] getent netgroup ngr1
>> >ngr1 (-,user1,ipa.test)
>> >[jhrozek@client] sssd $ [] getent netgroup ngr2
>> >[jhrozek@client] sssd $ []
>> >```
>> >
>> >After the patch:
>> >```
>> >[jhrozek@client] sssd $ [(review)] getent netgroup ngr1
>> >ngr1 (-,user1,ipa.test)
>> >[jhrozek@client] sssd $ [(review)] getent netgroup ngr2
>> >ngr2 (-,user1,ipa.test) (-,user1,ipa.test)
>> >[jhrozek@client] sssd $ [(review)]
>> >```
>> >
>> >So the netgroup can be resolved, but is it correct that the netgroup member
is listed twice?
>> >
>> It's impossible to say without content of netgroups on server.
>> Could you provide them?
>
>Same as the reproducer in the ticket. But the point is, is it ever OK to
>print duplicates? I guess they are harmless, but it just looks odd.
>
Then the bug is not fixed
Well, a different bug is fixed (and Michal was arguing there are two
bugs..). Before, the netgroup was not resolved at all, after the patch it is.
Here is a hierarchy based on descition in ticket
ng1: user1
ng2: user2, ng1
So getent netgroup shoudl return for
ng1 -> just a "(-,user1,-)"
ng2 -> "(-,user1,-)" "(-,user2,-)"
This is exactly a reason why I require to write a test for each bugfix.
Manual testing if error-prone and wasting of time.
Reviewer need a reliable reproducer and not just some "steps to reproduce"
OK, then my setup is different, sorry:
[jhrozek@unidirect] ~ $ [] ipa netgroup-show ngr1
Netgroup name: ngr1
NIS domain name: ipa.test
Member of netgroups: ngr2
Member User: user1
[jhrozek@unidirect] ~ $ [] ipa netgroup-show ngr2
Netgroup name: ngr2
NIS domain name: ipa.test
Member netgroups: ngr1
Member User: user1
"""
See the full comment at
https://github.com/SSSD/sssd/pull/59#issuecomment-255062876