[openstack-dev] [glance][tempest][defcore] Process to imrpove tests coverge in temepest
flavio at redhat.com
Wed Dec 9 14:02:53 UTC 2015
On 08/12/15 22:31 +0100, Jordan Pittier wrote:
>On Tue, Dec 8, 2015 at 9:52 PM, Flavio Percoco <flavio at redhat.com> wrote:
> Oh, I meant ocasionally. Whenever a missing test for an API is found,
> it'd be easy enough for the implementer to sohw up at the meeting and
> bring it up.
>From my experience as a Tempest reviewer, I'd say that most newly added tests
>are *not* submitted by Tempest regular contributors. I assume (wrongly ?) that
>it's mostly people from the actual projects (e.g glance) who are interested in
>adding new Tempest tests to test a feature recently implemented. Put
>differently, I don't think it's part of Tempest core team/community to add new
>tests. We mostly provide a framework and guidance these days.
I agree that the tempest team should focus on providing the framework
rather than the tests themselves. However, these tests are often
contributed by ppl that are not part of the project's team.
>But, reading this thread, I don"t know what to suggest. As a Tempest reviewer I
>won't start a new ML thread or send a message to a PTL each time I see a new
>test being added...I assume the patch author to know what he is doing, I can't
>keep on with what's going on in each and every project.
This is what I'd like to avoid. This assumption is exactly what almost
got the tasks API test almost merged and that will likely happen for
I don't think it's wrong to ping someone from the community when new
tests are added, especially because these tests are used by defcore
as well. Adding the PTL to the review (or some liaison) is simple
enough. We do this for many things in OpenStack. That is, we wait for
PTLs/liaisons approval before going forward with some decisions.
>Also, a test can be quickly removed if it is latter on deemed not so useful.
Sure but this is wasting people's time. The contributor's, reviewer's
and community's time as it'll have to be added, reviewed and then
I agree this doesn't happen too often but the fact that it happened is
enough of a reason for me to work on improving the process. Again,
especially because these tests are not meant to be used just by our
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 819 bytes
Desc: not available
More information about the OpenStack-dev