On Mon, 2009-07-20 at 14:39 +0200, Daniel Veillard wrote:
> On Mon, Jul 20, 2009 at 12:51:15PM +0100, Mark McLoughlin wrote:
> > --- a/src/qemu_conf.h
> > +++ b/src/qemu_conf.h
> > @@ -58,6 +58,11 @@ enum qemud_cmd_flags {
> > QEMUD_CMD_FLAG_KVM = (1 << 13), /* Whether KVM is compiled in */
> > QEMUD_CMD_FLAG_DRIVE_FORMAT = (1 << 14), /* Is -drive format= avail */
> > QEMUD_CMD_FLAG_VGA = (1 << 15), /* Is -vga avail */
> > +
> > + /* features added in qemu-0.10.0 */
> > + QEMUD_CMD_FLAG_0_10 = (1 << 16),
> > + QEMUD_CMD_FLAG_NET_NAME = QEMUD_CMD_FLAG_0_10, /* -net ...,name=str */
> > + QEMUD_CMD_FLAG_HOST_NET_ADD = QEMUD_CMD_FLAG_0_10, /* host_net_add monitor command */
> > };
>
> Hum, defining multiple time the same value in an enum, maybe that's
> fine but that looks weird to me, especially as each entry so far was
> about separated capabilities, independantly of the potential version.
>
> Not a big deal but what do others think ?
Well my thinking was:
- We can't easily probe for the monitor command without a bunch of
code
- The name param was only introduced in 0.10
- You need both for nic hotplug
- Parsing 'qemu -help' sucks and qemu has a much saner release cycle
now, so relying on version numbers makes more sense
- The FLAG_0_10 thing is there mostly as documentation and we can
easily split it into two flags if we need to in future
But agree it's not a big deal - willing to do whatever I'm told to here
and I'm guessing danpb has a firm opinion on it :-)
Cheers,
Mark.
--
Libvir-list mailing list
Libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list