[Gluster-devel] Should we remove tests/basic/rpm.t from the regression suite?

[Kaushal M]

On Tue, Dec 8, 2015 at 9:39 PM, Kaleb KEITHLEY <kkeithle at redhat.com> wrote:
> On 12/08/2015 10:25 AM, Kaushal M wrote:
>> We already have devrpms jobs which test building rpms, so building
>> rpms again seems unnecessary.
>>
>> The reason I'm asking this is because I'm facing issues trying to get
>> this working in the saltstack configured environment I'm fixing. I
>> don't see what's specifically wrong with the test, but it just doesn't
>> seem to work. So, I was wondering why even run it.
>
> IIRC, it's just to test that glusterfs.spec(.in) didn't break.
>
> Is it the devrpm job's job to build rpms for the following tests? Or is
> it actually a test that the .spec didn't break? Or is the implicit test
> that the .spec didn't break just a side-effect? And does it matter
> which?  ;-)

It shouldn't matter. RPMs being built means the spec didn't break.
Shouldn't matter if the RPMs are used for other tests. Right now RPMs
being built by devrpms isn't being used for any other tests, so their
only purpose is to test the spec.

>
> --
>
> Kaleb
>