Comment 5 for bug 1866106

Yeah ... I'm trying to think what to do here because this CONF.pointer_model option is supposed to be available for all virt drivers ... it's just only used in the libvirt driver.

Maybe it should have been defined strictly under CONF.libvirt.pointer_model, then it would be more straightforward what to do with it.

If None worked, None and ps2mouse are doing the same thing today, they're exposing the default behavior. You're not actually "choosing" anything when you choose 'ps2mouse'. So I think that's misleading too.

So ... the easiest thing would be to use the proposed patch as-is, to simply fix the treatment of the intended None choice. That would punt the misleading nature of the options.

But if we re-evaluate all the choices, it seems to me the most valid/straightforward choices would be something like: 'usbtablet' and 'default', where 'default' uses default behavior.

I could update my patch to add a deprecation note for 'ps2mouse' and add 'default' to replace it, and then remove the broken None option altogether.