Re: [Qemu-devel] [PATCH 1/2] tests: Test case for object_resolve_path*()

[Eduardo Habkost <ehabkost@redhat.com>]

On Mon, Jul 10, 2017 at 10:10:41AM +0200, Igor Mammedov wrote:
> On Fri,  7 Jul 2017 18:30:51 -0300
> Eduardo Habkost <ehabkost@redhat.com> wrote:
> 
> > Test for partial path lookup using object_resolve_path*().
> > 
> > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> > ---
> >  tests/check-qom-proplist.c | 41 +++++++++++++++++++++++++++++++++++++++++
> >  1 file changed, 41 insertions(+)
> > 
> > diff --git a/tests/check-qom-proplist.c b/tests/check-qom-proplist.c
> > index 8e432e9..abafbd7 100644
> > --- a/tests/check-qom-proplist.c
> > +++ b/tests/check-qom-proplist.c
> > @@ -568,6 +568,46 @@ static void test_dummy_delchild(void)
> >      object_unparent(OBJECT(dev));
> >  }
> >  
> > +static void test_qom_partial_path(void)
> > +{
> > +    Object *root  = object_get_objects_root();
> > +    Object *cont1 = container_get(root, "/cont1");
> > +    Object *obj1  = object_new(TYPE_DUMMY);
> > +    Object *obj2a = object_new(TYPE_DUMMY);
> > +    Object *obj2b = object_new(TYPE_DUMMY);
> > +    bool ambiguous;
> > +
> > +    /* Objects created:
> > +     * /cont1
> > +     * /cont1/obj1
> > +     * /cont1/obj2 (obj2a)
> > +     * /obj2 (obj2b)
> > +     */
> > +    object_property_add_child(cont1, "obj1", obj1, &error_abort);
> > +    object_unref(obj1);
> > +    object_property_add_child(cont1, "obj2", obj2a, &error_abort);
> > +    object_unref(obj2a);
> > +    object_property_add_child(root,  "obj2", obj2b, &error_abort);
> > +    object_unref(obj2b);
> > +
> > +    ambiguous = false;
> > +    g_assert(!object_resolve_path_type("", TYPE_DUMMY, &ambiguous));
> > +    g_assert(ambiguous);
> > +
> > +    ambiguous = false;
> > +    g_assert(!object_resolve_path("obj2", &ambiguous));
> > +    g_assert(ambiguous);
> > +
> > +    ambiguous = false;
> > +    g_assert(object_resolve_path("obj1", &ambiguous) == obj1);
> > +    g_assert(!ambiguous);
> 
> I'd also add test case for
>   object_resolve_path(..., NULL)

They are added by patch 2/2, after the bug is fixed.

> 
> > +
> > +    object_unparent(obj1);
> > +    object_unparent(obj2a);
> > +    object_unparent(obj2b);
> Are above unparenting is necessary?
> 
> > +    object_unparent(cont1);
> Wouldn't parent destruction sufficient to trigger
> implicit destruction of children?

Probably it is.  I will test it.

-- 
Eduardo