On Sun, 2017-01-01 at 12:35 +0400, Roman Bogorodskiy wrote:
> No rush, I see Michal has yet another proposal for this that we
haven't
> considered and even though there are somedrawbacks to that as well, it
> looks nicer than this.
>
> After all the ideas I'm starting to like the "gross" one the
best. Oh
> my =)
My vote still goes to this solution, because having a placeholder seems
more explicit and easier to follow than doing s/lo/lo0/ directly.
I vote for Michal's approach as it doesn't require us to
disable VIR_TEST_REGENERATE_OUTPUT.
--
Andrea Bolognani / Red Hat / Virtualization