[libvirt] [PATCH 1/2] qemu: Invert condition nesting in qemuDomainDefValidate()

[Andrea Bolognani]

On Fri, 2017-12-15 at 14:20 +0100, Peter Krempa wrote:
> > > You know that generally we prefer variables declared at the top scope?
> > 
> > Is that so? I've seen several instances of the opposite, and it
> > makes to me not to pollute the function scope with one-use
> > variables when there are usually enough variables that actually
> > need to be accessed throughout the function. But I can leave the
> > declaration where it is if you like it better that way.
> 
> Fair enough. We only state that we should declare it at the beginning of
> a scope, so while generally most are declared at function scoep, this
> does not violate the contributor guidelines.

Cool, I'll still move it to the inner scope then.

> > It's very simple, anyway: the version number was wrong, since
> > QEMU introduced query-hotpluggable-cpus in 2.7 rather than 2.5.
> 
> While this is true for the presence of query-hotpluggable-cpus. The
> paragraph is specifically saying that QEMU rejects such configurations
> starting from 2.5., but the code checks them only since
> query-hotpluggable-cpus (which was introduced in 2.7).
> 
> Your modification is thus incorrect. You can add statement that
> query-hotpluggable-cpus was added in 2.7. but mangling it as you did is
> not correct in my opinion.

I see, that's a nuance I was not aware of. I'll amend the comment
so that it tells the whole story.

-- 
Andrea Bolognani / Red Hat / Virtualization