[PATCH 2/2] tests: test the wl_display_roundtrip_queue() function

Marek Chalupa mchqwerty at gmail.com
Thu Aug 21 02:05:24 PDT 2014


The issues I pointed out are not anything big. I'll send a patch that
removes wl_display_dispatch_pending from both test-cases and I noticed that
there are more leaks it the queue-test, so I'll get rid of them all in
another patch if it's OK :)

Thanks,
Marek



On 21 August 2014 10:57, Pekka Paalanen <ppaalanen at gmail.com> wrote:

> On Thu, 21 Aug 2014 10:46:55 +0200
> Marek Chalupa <mchqwerty at gmail.com> wrote:
>
> > Hi,
> >
> >
> > On 20 August 2014 18:29, Giulio Camuffo <giuliocamuffo at gmail.com> wrote:
> >
> > > ---
> > >  tests/queue-test.c | 62
> > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> > >  1 file changed, 62 insertions(+)
> > >
> > > diff --git a/tests/queue-test.c b/tests/queue-test.c
> > > index a4b165d..fc8a920 100644
> > > --- a/tests/queue-test.c
> > > +++ b/tests/queue-test.c
> > > @@ -158,6 +158,63 @@ client_test_multiple_queues(void)
> > >  }
> > >
> > >  static void
> > > +sync_callback_roundtrip(void *data, struct wl_callback *callback,
> > > uint32_t serial)
> > > +{
> > > +       bool *done = data;
> > > +       *done = true;
> > > +}
> > > +
> > > +static const struct wl_callback_listener sync_listener_roundtrip = {
> > > +       sync_callback_roundtrip
> > > +};
> > > +
> > > +/* Test that doing a roundtrip on a queue only the events on that
> > > + * queue get dispatched. */
> > > +static int
> > > +client_test_queue_roundtrip(void)
> > > +{
> > > +       struct wl_event_queue *queue;
> > > +       struct wl_callback *callback1;
> > > +       struct wl_callback *callback2;
> > > +       struct wl_display *display;
> > > +
> > > +       display = wl_display_connect(SOCKET_NAME);
> > > +       client_assert(display);
> > > +
> > > +       /* Make the current thread the display thread. This is because
> > > +        * wl_display_dispatch_queue() will only read the display fd if
> > > +        * the main display thread has been set. */
> > > +       wl_display_dispatch_pending(display);
> > >
> >
> > I don't think this is necessary,  this behavior was introduced in
> >  385fe30e8b144a968aa88c6546c2ef247771b3d7 and
> > 78cfa967681c965d23f6cbf76e080bbb0b564ff6
> > and it was removed
> > in 3c7e8bfbb4745315b7bcbf69fa746c3d6718c305 (now there's only unused
> > variable display_thread in wl_display structure, we were discussing it
> back
> > on the list IIRC).
> >
> > >
> > > +       queue = wl_display_create_queue(display);
> > > +       client_assert(queue);
> > > +
> > > +       bool done1 = false;
> > > +       callback1 = wl_display_sync(display);
> > > +       assert(callback1 != NULL);
> > > +       wl_callback_add_listener(callback1, &sync_listener_roundtrip,
> > > &done1);
> > > +
> > > +       bool done2 = false;
> > > +       callback2 = wl_display_sync(display);
> > > +       assert(callback2 != NULL);
> > > +       wl_callback_add_listener(callback2, &sync_listener_roundtrip,
> > > &done2);
> > > +       wl_proxy_set_queue((struct wl_proxy *) callback2, queue);
> > > +
> > > +       wl_display_roundtrip(display);
> > > +       client_assert(done1 == true);
> > > +       client_assert(done2 == false);
> > > +
> > > +       wl_display_roundtrip_queue(display, queue);
> > > +       client_assert(done2 == true);
> > > +
> > > +       wl_callback_destroy(callback1);
> > > +       wl_callback_destroy(callback2);
> > >
> >
> > You're leaking the queue here.
> >
> >
> > > +       wl_display_disconnect(display);
> > > +
> > > +       return 0;
> > > +}
> > > +
> > > +static void
> > >  client_alarm_handler(int sig)
> > >  {
> > >         fprintf(stderr, "Received SIGALRM signal, aborting.\n");
> > > @@ -199,6 +256,11 @@ client_main(int fd)
> > >                 return EXIT_FAILURE;
> > >         }
> > >
> > > +       if (client_test_queue_roundtrip() != 0) {
> > > +               fprintf(stderr, "queue rountrip test failed\n");
> > > +               return EXIT_FAILURE;
> > > +       }
> > > +
> > >         return EXIT_SUCCESS;
> > >  }
> > >
> > > --
> > > 2.0.4
> > >
> > > _______________________________________________
> > > wayland-devel mailing list
> > > wayland-devel at lists.freedesktop.org
> > > http://lists.freedesktop.org/mailman/listinfo/wayland-devel
> > >
> >
> > Otherwise:
> >
> > Reviewed-by: Marek Chalupa <mchqwerty at gmail.com>
>
> Oops, just pushed it already. Anyone care to follow up? :-)
>
>
> Thanks,
> pq
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/wayland-devel/attachments/20140821/941ea657/attachment.html>


More information about the wayland-devel mailing list