Daniel P. Berrangé <berrange@xxxxxxxxxx> writes: > On Tue, Oct 26, 2021 at 05:15:10PM +0200, Markus Armbruster wrote: >> Daniel P. Berrangé <berrange@xxxxxxxxxx> writes: >> >> > On Tue, Oct 26, 2021 at 11:37:19AM +0200, Markus Armbruster wrote: [...] >> >> Management applications are better off with a feature flag than with a >> >> naming convention we sometimes ignore. >> > >> > We will sometimes ignore/forget the feature flag too though, so I'm >> > not convinced there's much difference there. >> >> -compat unstable-input=reject,unstable-output=hide should help you stay >> on the straight & narrow :) > > That's from the pov of the mgmt app. I meant from the POV of QEMU > maintainers forgetting to add "unstable" flag, just as they might > forget to add a "x-" prefix. Got it. My point was that feature flag "unstable" is an unequivocal signal for "this thing is unstable", while a name starting with "x-" isn't: there are exceptions. The converse is a wash: we can forget to mark something unstable no matter how the mark works.