On Mon, Dec 14, 2015 at 08:55:01AM +0200, Liran Schour wrote:
> Andy Zhou <az...@ovn.org> wrote on 12/12/2015 12:37:32 AM:
>
> > On Tue, Dec 1, 2015 at 6:20 AM, Liran Schour <lir...@il.ibm.com> wrote:
> > "dev" <dev-boun...@openvswitch.org> wrote on 25/11/2015 12:16:01 AM:
> >
> > > From: Andy Zhou <az...@nicira.com>
> > > @@ -617,6 +621,101 @@ monitor_print(struct json *table_updates,
> > > }
> > >
> > > static void
> > > +monitor2_print_row(struct json *row, const char *type, const char
> > *uuid,
> > > + const struct ovsdb_column_set *columns, struct
> table
> > *t)
> > > +{
> > > + if (!strcmp(type, "delete")) {
> > > + if (row->type != JSON_NULL) {
> > > + ovs_error(0, "delete method does not expect <row>");
> > > + return;
> > > + }
> > > +
> > > + table_add_row(t);
> > > + table_add_cell(t)->text = xstrdup(uuid);
> > > + table_add_cell(t)->text = xstrdup(type);
> > > + } else {
> > > + if (!row || row->type != JSON_OBJECT) {
> > > + ovs_error(0, "<row> is not object");
> > > + return;
> > > + }
> > > + monitor_print_row(row, type, uuid, columns, t);
> > > + }
> > > +}
> >
> > Update2 does not send default values, therefore we will miss these
> values
> > on ovsdb-client.
> > I can fix it on monitor_cond or you can send an updated patch.
> >
> > What changes do you have in mind?
>
> > The goal of "monitor" command is to show what's being transmitted
> > per ovsdb remote protocol. Since
> > the default values are not being transmitted, I'd think it is O.K.
> > not to show them. No?
> >
>
> OK. I thought the user expects to see the real outcome of the updates. If
> not and "monitor" command should show only what is being transmitted it is
> OK like it is now.
I guess that the purpose could be argued either way, but I think it's
reasonable to have it work the way Andy did it.
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
