[SSSD] [PATCH] sdap_fill_memberships: continue if a member is not foud in sysdb

Jakub Hrozek jhrozek at redhat.com
Tue Feb 26 23:21:57 UTC 2013


On Mon, Feb 25, 2013 at 11:23:12AM +0100, Pavel Březina wrote:
> On 02/22/2013 05:57 PM, Jakub Hrozek wrote:
> >On Fri, Feb 22, 2013 at 03:32:50PM +0100, Pavel Březina wrote:
> >>https://fedorahosted.org/sssd/ticket/1755
> >
> >I agree, I can't think of a reason why we should be failing to store a
> >group completely when we can't find one of its members for one reason or
> >another.
> >
> >I haven't really done much testing but ack on the approach.
> >

Now I did. Ack.

> >I think you forgot to attach the unit test, though. Or is that coming
> >together with the refactoring patches?
> 
> This is not part of the refactoring so I didn't plan to unit test
> it. I want to test only sdap_nested_group_send() (from the
> refactoring patch).
> 

Fair enough. I'm looking forward to the tests being part of the nested
group processing refactor, then.

> I think we should create separate tickets to cover providers with
> unit tests.

In general, sure, we might even have a new developer joining us for the
Google Summer of code to cover the SSSD better with unit tests.

But each new feature or big refactor should be covered with tests from
now on. I don't think we can afford the breakage we had around 1.9 again.
It was too costly in devel time and could have been caught much sooner
with better testing.



More information about the sssd-devel mailing list