On Mon, Jun 25, 2012 at 12:37:48PM +0100, Daniel P. Berrange wrote:
On Mon, Jun 25, 2012 at 01:34:41PM +0200, Christophe Fergeau wrote:
> On Sat, Jun 23, 2012 at 10:17:57PM +0300, Zeeshan Ali (Khattak) wrote:
> > From: "Zeeshan Ali (Khattak)" <zeeshanak(a)gnome.org>
> >
> > - gvir_config_capabilities_cpu_add_feature
> > - gvir_config_capabilities_cpu_set_topology
>
> As I understand it, the capabilities XML is read-only, the only reason for
> these setters to exist is for use with GVirConfigDomainCpu which is added
> later in this patch set. I think I'd put the setters in this new class even
> if this makes the API asymetric (getters in one class, setters in another).
Yes & no. Technically you are correct that the capabilities XML is
readonly, however, when using the test driver in libvirt, it is
possible to pass a custom XML document fully describing the capabilities
that the test driver should return. Thus it is useful to have setters
in the GVirConfigCapabilities class
OK, thanks for the clarification!
Christophe