Open Stack

Sean, Giulio,

Thanks a lot for your explanations.

Giulio, from my opinion, as you said, the second solution would probably be
better (the two backend names at the class level and skip the test if
name1=name2). Let's see what others think about that.

Jérôme


On Fri, May 31, 2013 at 9:38 PM, Giulio Fidente <gfidente at redhat.com> wrote:

> On 05/31/2013 09:23 PM, Giulio Fidente wrote:
>
>> On 05/31/2013 07:07 PM, Sean Dague wrote:
>>
>>> This is why changing the variable later isn't any good (in the second
>>> patch), because the decorator was already evaluated. You may be able to
>>> build a trickier decorator which does a late load based on the class.
>>>
>>
>> got it
>>
>> so would you prefer (and eventually +1) to evaluate the boolean within
>> the test case or rather move at the class level the two backend names
>> and skipIf name1=name2 '
>>
>
> I think the latter would work better but this raises another question:
> misconfiguring tempest by setting two backends with same name could result
> in a false positive
>
> shall we skip if misconfigured or what?
>
> --
> Giulio Fidente
> GPG KEY: 08D733BA | IRC: giulivo
>
> ______________________________**_________________
> openstack-qa mailing list
> openstack-qa at lists.openstack.**org <openstack-qa at lists.openstack.org>
> http://lists.openstack.org/**cgi-bin/mailman/listinfo/**openstack-qa<http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-qa>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-qa/attachments/20130603/a306e64d/attachment.html>