[libvirt] PATCH: 3/5: The easy Xen and QEMU implementations

[Daniel Veillard]

On Wed, May 13, 2009 at 05:08:00PM +0100, Daniel P. Berrange wrote:
> > > +cleanup:
> > > +    for (tmp = retargv ; tmp && *tmp ; tmp++)
> > 
> >   weird, we use a static string just for "" ?
> > 
> > > +        VIR_FREE(*tmp);
> > > +    VIR_FREE(retargv);
> > > +
> > > +    for (tmp = retenv ; tmp && *tmp ; tmp++)
> > 
> >   same
> 
> I'm not sure what you mean here ?

  I got confused about the 'tmp && *tmp' 
Somehow my brain expect the tmp check to be out of the loop since
the step 'tmp++' should not get it back to NULL.

> > > --- a/src/xen_unified.h	Wed May 13 13:13:18 2009 +0100
> > > +++ b/src/xen_unified.h	Wed May 13 13:16:50 2009 +0100
> > > @@ -46,6 +46,9 @@ extern int xenRegister (void);
> > >  
> > >  #define MIN_XEN_GUEST_SIZE 64  /* 64 megabytes */
> > >  
> > > +#define XEN_CONFIG_FORMAT_XM    "xen-xm"
> > > +#define XEN_CONFIG_FORMAT_SEXPR "xen-sxpr"
> > 
> >   Actually, shouldn't those be part of the public headers instead
> > I'm afraid I'm missing something here ...
> 
> You'd pass these kind of values into the 'char *nativeFormat' parameter
> but I wasn't really considering them part of the stable ABI/API. The
> config formats each driver supports are specific to particular backends
> and particular HV versions. So I decided it was better not to expose
> them directly.  Perhaps we could add them to the capabilities XML
> format though, as dynamically exported data, so they're not ABI.

  well that would be fine too, just that user shouldn't have to google
  or guess to make sense of it.

Daniel

-- 
Daniel Veillard      | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
daniel at veillard.com  | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library  http://libvirt.org/