Re: [Xen-devel] [PATCH, v3]: xl: randomly generate UUID's

[Gianni Tedesco <gianni.tedesco@xxxxxxxxxx>]

On Fri, 2010-08-27 at 16:13 +0100, Christoph Egger wrote:
> On Friday 27 August 2010 15:56:35 Gianni Tedesco wrote:
> > On Thu, 2010-08-26 at 16:41 +0100, Christoph Egger wrote:
> > > @@ -2219,8 +2224,11 @@ static void list_domains(int verbose, co
> > >                  info[i].dying ? 'd' : '-',
> > >                  ((float)info[i].cpu_time / 1e9));
> > >          free(domname);
> > > -        if (verbose)
> > > -            printf(" " LIBXL_UUID_FMT, LIBXL_UUID_BYTES(info[i].uuid));
> > > +        if (verbose) {
> > > +            char buf[LIBXL_UUID_LEN];
> > > +            libxl_uuido_string(&info[i].uuid, buf, sizeof(buf));
> > > +            printf(" %s", buf);
> > > +        }
> > >          putchar('\n');
> > >      }
> > >  }
> >
> > Really? There's no way to keep the macro approach and get at the thing
> > like an array of bytes?
> 
> Yes. gcc complains about 'subscripted value is neither array nor pointer'.

Yes, my implementation may have been broken but was there a way to make
it work? I assume there must be. Basically this allows us to avoid the
use of a temporary buffer when printing UUID strings.

> > If that's the case then how can libxl_uuid_bytearray() that I introduced
> > work correctly? 
> 
> I removed it because its redundant to libxl_uuido_string().
> A bytearray and a string is the same in C unless you use unicode.

I think we're talking accross purposes, the point of the
libxl_uuid_bytearray is basically to cast the platform specific type to
a uint8_t * so that we can access the uuid as a 'string' - this is
different to a textual representation of a UUID.

I would be willing to bet money that this can always be done since it's
the only reasonable way to represent a UUID internally.

Gianni


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel