Eric Blake <ebl...@redhat.com> writes: > Right now, qmp-output-visitor happens to produce a QNull result > if nothing is actually visited between the creation of the visitor > and the request for the resulting QObject. A stronger protocol > would require that a QMP output visit MUST visit something. But > to still be able to produce a JSON 'null' output, we need a new > visitor function that states our intentions.
Overdue. When we extended the json-parser to accept null, we neglected to extend visitors accordingly. We need to: * Extend the visitor core (this patch). * Implement it in at least the visitors that aren't restricted to a subset: dealloc (this patch), QMP input (next patch), QMP output (patch after next, together with other stuff). * Update users, if any. * If QAPI had a 'null' type, we'd also have to use it in the generated visitor functions. It doesn't. I think the 'any' type can hold a null, but that's it. A 'null' type might be useful as an alternate member type. We'll create one when we need it. > This patch introduces the new visit_type_null() interface, and > later patches will then wire it up into the qmp visitors. > > Signed-off-by: Eric Blake <ebl...@redhat.com> > Reviewed-by: Marc-André Lureau <marcandre.lur...@redhat.com> > > --- > v9: no change > v8: rebase to 'name' motion > v7: new patch, based on discussion about spapr_drc.c > --- > include/qapi/visitor-impl.h | 3 +++ > include/qapi/visitor.h | 8 ++++++++ > qapi/qapi-dealloc-visitor.c | 5 +++++ > qapi/qapi-visit-core.c | 5 +++++ > 4 files changed, 21 insertions(+) > > diff --git a/include/qapi/visitor-impl.h b/include/qapi/visitor-impl.h > index aab46bc..8705136 100644 > --- a/include/qapi/visitor-impl.h > +++ b/include/qapi/visitor-impl.h > @@ -75,6 +75,9 @@ struct Visitor > * visitors do not currently visit arbitrary types). */ > void (*type_any)(Visitor *v, const char *name, QObject **obj, > Error **errp); > + /* Must be provided to visit explicit null values (right now, only the > + * dealloc visitor supports this). */ Will need updating to match whatever convention we pick in the previous patch for documenting "mandatory to visit X", and the restrictions on visitor use resulting from some of them not implementing it. > + void (*type_null)(Visitor *v, const char *name, Error **errp); > > /* May be NULL; most useful for input visitors. */ > void (*optional)(Visitor *v, const char *name, bool *present); [...]
