On Wed, 2016-01-06 at 10:27 +0100, Wido den Hollander wrote:
On 01/06/2016 09:49 AM, Andrea Bolognani wrote:
> On Wed, 2016-01-06 at 09:32 +0100, Wido den Hollander wrote:
> > diff --git a/src/util/virstoragefile.c b/src/util/virstoragefile.c
> > index 2aa1d90..1354601 100644
> > --- a/src/util/virstoragefile.c
> > +++ b/src/util/virstoragefile.c
> > @@ -2246,7 +2246,7 @@ virStorageSourceRBDAddHost(virStorageSourcePtr src,
> > if (VIR_STRDUP(src->hosts[src->nhosts - 1].port, port) < 0)
> > goto error;
> > } else {
> > - if (VIR_STRDUP(src->hosts[src->nhosts - 1].port,
"6789") < 0)
> > + if (VIR_STRDUP(src->hosts[src->nhosts - 1].port, NULL) < 0)
> > goto error;
> > }
>
> Shouldn't this be
>
> > if (VIR_STRDUP(src->hosts[src->nhosts - 1].port, port) < 0)
> > goto error;
> > } else {
> > - if (VIR_STRDUP(src->hosts[src->nhosts - 1].port,
"6789") < 0)
> > - goto error;
> > + src->hosts[src->nhosts - 1].port = NULL;
> > }
What does that benefit? We set it to 6789 and to NULL afterwards? I just
want to get rid of any hardcoded monitor ports in libvirt for Ceph.
Look again, you'll notice it's a diff :)
I still set the port to NULL, just like your code does, but instead of
using VIR_STRDUP() I use plain assignment.
Cheers.
--
Andrea Bolognani
Software Engineer - Virtualization Team