On Mon, 2015-11-16 at 13:16 +0000, Paul Durrant wrote:
> >
> > > + ret = vasprintf(&s, fmt, ap);
> > > if (ret == -1) {
> > > return -1;
> > > }
> > > xs_write(ctx->xsh, t, path, s, ret);
> > > + if (perms)
> > > + xs_set_permissions(ctx->xsh, t, path, perms, num_perms);
> >
> > This can fail, can't it? (OTOH so can xs_write, so maybe there is some
> > reason we apparently don't care for such things here?)
> >
>
> That's a good point. I would have thought wiring an xs_write() failure
> back through to the caller would be a good idea (and same goes for
> xs_set_permissions, I guess).
git tells me it has been this way since libxl was first committed to the
repo. I can't think of a good reason to squash these errors.
Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
