Re: [PATCH v7 3/8] Acceptance tests: use avocado tags for machine type

[Cleber Rosa <crosa@redhat.com>]

On Fri, Nov 08, 2019 at 02:20:45PM +0100, Philippe Mathieu-Daudé wrote:
> On 11/4/19 4:13 PM, Cleber Rosa wrote:
> >           """
> > -        self.vm.set_machine('none')
> >           self.vm.add_args('-S')
> >           self.vm.launch()
> > diff --git a/tests/acceptance/linux_initrd.py b/tests/acceptance/linux_initrd.py
> > index c61d9826a4..3a0ff7b098 100644
> > --- a/tests/acceptance/linux_initrd.py
> > +++ b/tests/acceptance/linux_initrd.py
> > @@ -20,6 +20,7 @@ class LinuxInitrd(Test):
> >       Checks QEMU evaluates correctly the initrd file passed as -initrd option.
> >       :avocado: tags=arch:x86_64
> > +    :avocado: tags=machine:pc
> 
> For some tests we can run on multiple machines (here q35), I was tempted to
> use multiple tags. How could I do that now?
>

I missed this comment: you can add many tag values here to *classify*
the test as being "q35 machine type capable".

But, Avocado will only run a test multiple times with a varianter
plugin active.  In that case, a "machine" *parameter* with different
values will be passed to the tests.  This tag value is being used
as a default value for the parameter, so it has a lower precedence.

We have a pending task[1] to create an initial CIT file for arch and
machine types.

CC'ing Jan Richter, who is supposed to start working on it soon.

- Cleber.

[1] - https://trello.com/c/1wvzcxHY/105-create-cit-parameter-for-acceptance-tests