Re: [Qemu-devel] Proposal for extensions of block job commands in QEMU 1.2

[Anthony Liguori]

On 05/21/2012 09:16 AM, Luiz Capitulino wrote:
> On Mon, 21 May 2012 09:10:40 -0500
> Anthony Liguori<address@hidden>  wrote:
>
> > On 05/21/2012 08:59 AM, Luiz Capitulino wrote:
> > > On Fri, 18 May 2012 19:08:42 +0200
> > > Paolo Bonzini<address@hidden>   wrote:
> > >
> > > > Modified QMP commands
> > > > =====================
> > >
> > > As we have discussed on the ML, we're not going to extend QMP commands.
> > >
> > > I understand your reasoning, and since the beginning I thought this was
> > > something useful to do, but we've already settled for not doing this.
> > >
> > > I also think that we shouldn't have exceptions, as in practice this means
> > > we're extending commands anyway. So either, we do it or we don't.
> >
> > Well, I think we should ask ourselves the following question:
> >
> > How would a client figure out if the new options are available?
> >
> > This is the primary reason for not extending existing commands.
>
> Yes, I know. But if Paolo implements schema introspection, would you
> agree on extending commands.
>
> You seemed to be against even if we had schema introspection.

I'm not against it in principle, just in practice.  Today, checking whether a 
command exists is:

commands = qmp.query_commands()

if 'block-stream' in commands:
    # has block-stream

I have a hard time envisioning how schema introspection can be reasonably 
implemented in a client.

If we really feel it's important to extend commands, I'd prefer something like 
per-command capabilities.

Regards,

Anthony Liguori