[ovirt-devel] vdsm make rpm & make check on fe vs el

[Nir Soffer]

----- Original Message -----
> From: "Antoni Segura Puimedon" <asegurap at redhat.com>
> To: "Francesco Romani" <fromani at redhat.com>
> Cc: devel at ovirt.org
> Sent: Monday, October 27, 2014 1:26:47 PM
> Subject: Re: [ovirt-devel] vdsm make rpm & make check on fe vs el
> 
> 
> 
> ----- Original Message -----
> > From: "Francesco Romani" <fromani at redhat.com>
> > To: devel at ovirt.org
> > Sent: Monday, October 27, 2014 12:20:17 PM
> > Subject: Re: [ovirt-devel] vdsm make rpm & make check on fe vs el
> > 
> > ----- Original Message -----
> > > From: "Nir Soffer" <nsoffer at redhat.com>
> > > To: "Francesco Romani" <fromani at redhat.com>
> > > Cc: devel at ovirt.org, "Mooli Tayer" <mtayer at redhat.com>
> > > Sent: Monday, October 27, 2014 12:16:11 PM
> > > Subject: Re: [ovirt-devel] vdsm make rpm & make check on fe vs el
> > 
> > 
> > [...]
> > > > > > I've also saw a different numbers of tests running for make rpm on
> > > > > > el.
> > > > > 
> > > > > I have seen this on several 6.5 machines - make rpm on does *not* run
> > > > > the tests.
> > > > > 
> > > > > It would be nice if take a look at this.
> > > > 
> > > > It was made on purposes:
> > > > 
> > > > # Skips check since rhel default repos lack pep8 and pyflakes
> > > > %if ! 0%{?rhel}
> > > > %global with_check 1
> > > > %endif
> > > > 
> > > > http://gerrit.ovirt.org/#/c/29213/
> > > > 
> > > > I believe time has come for a more robust solution
> > > 
> > > As first aid, we should log something like "Skipping tests because ..."
> > > in this case.
> > > 
> > > I think we need to change that so we disable only pep8 in make rpm.
> > > 
> > > Nir
> > 
> > +1
> > 
> > we can afford to skip local checks (pep8/pyflakes), but the core test suite
> > should run
> > anyway.
> > 
> > I'll post a patch today(ish) if noone is faster/disagrees with the above :)
> 
> I don't think the tests should run on make rpm. In make check, sure. In make
> rpm...
> It's a last resort for making people realize something is broken when they
> are not
> running make check.

In a perfect world, make rpm would only make rpm and we would be happy.

But in our world, the tests run in make rpm may fail even if they
succeed in make check. This may happen if you forgot to update the 
makefiles after adding new test modules or libraries.

So the safest way to make sure that the test would run from source rpm
is to run them in make rpm.

Nir