On Fri, Nov 13, 2026 at 11:14 AM Markus Armbruster <[email protected]> wrote:
> David wants to do the same but with AI.  He "pre-arranges", and Bob
> agrees to review.  Does Charlie get a vote?  What about me?

Charlie the co-maintainer does "get a vote" on merging, yes, and she
is free to leave the review to Bob as well. Well, technically we do
not operate by voting, but co-maintainers are supposed to operate by
consensus and this is no exception.

You the QMP maintainer... it depends. Probably not if the only change
is along the lines of:

-{ 'enum': 'Accelerator', 'data': ['hvf', 'kvm', 'nvmm', 'qtest',
'tcg', 'whpx', 'xen'] }
+{ 'enum': 'Accelerator', 'data': ['hvf', 'kvm', 'mshv', 'nvmm',
'qtest', 'tcg', 'whpx', 'xen'] }

Even larger QMP interfaces may fall under bullet #3 - parts of large
code changes - but you definitely get a vote if there are substantial
changes to code you maintain. No doubt about that! This is just normal
interaction between maintainers; I think we do well in that respect
and I see no reason for LLMs to change that.

We may have to become more accustomed to saying no to contributors and
even co-maintainers, but I think we're good at building/accepting
consensus. I'm sure that stuff like the Meson migration was seen with
skepticism by some members of the community, but it was accepted
nevertheless. (I hope it has proven itself; and I was also on the
other side of the fence with Pierrick's recent tests/tcg overhaul,
btw).

Paolo

> Charlie and I can of course leave the review to Bob.  But we'll have to
> live with whatever makes it into the tree.
>
> I fear this risks creating resentment among maintainers.
>
> [...]
>


Reply via email to