[PATCH libevdev 9/9] test: test for logging function

[Peter Hutterer]

On Mon, Sep 09, 2013 at 05:30:29PM +0200, David Herrmann wrote:
> Hi Benjamin
> 
> On Mon, Sep 9, 2013 at 5:09 PM, Benjamin Tissoires
> <benjamin.tissoires at gmail.com> wrote:
> > On Tue, Sep 3, 2013 at 9:07 AM, Peter Hutterer <peter.hutterer at who-t.net> wrote:
> >> Signed-off-by: Peter Hutterer <peter.hutterer at who-t.net>
> >> ---
> >>  test/test-libevdev-init.c | 25 +++++++++++++++++--------
> >>  1 file changed, 17 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/test/test-libevdev-init.c b/test/test-libevdev-init.c
> >> index 8f1d0d1..2610975 100644
> >> --- a/test/test-libevdev-init.c
> >> +++ b/test/test-libevdev-init.c
> >> @@ -93,10 +93,16 @@ START_TEST(test_init_and_change_fd)
> >>  }
> >>  END_TEST
> >>
> >> +
> >> +static int log_fn_called = 0;
> >> +static char *logdata = "test";
> >>  static void logfunc(enum libevdev_log_priority priority,
> >>                     void *data,
> >>                     const char *file, int line, const char *func,
> >> -                   const char *f, va_list args) {}
> >> +                   const char *f, va_list args) {
> >> +       ck_assert_int_eq(strcmp(logdata, data), 0);
> >> +       log_fn_called++;
> >> +}
> >>
> >>  START_TEST(test_log_init)
> >>  {
> >> @@ -107,16 +113,19 @@ START_TEST(test_log_init)
> >>
> >>         dev = libevdev_new();
> >>         ck_assert(dev != NULL);
> >> -       libevdev_set_log_function(logfunc, NULL);
> >> -       libevdev_free(dev);
> >>
> >> -       dev = libevdev_new();
> >> -       ck_assert(dev != NULL);
> >> +       libevdev_set_log_function(logfunc, logdata);
> >> +       libevdev_next_event(dev, LIBEVDEV_READ_NORMAL, NULL);
> >> +
> >>         libevdev_set_log_function(NULL, NULL);
> >> -       libevdev_set_log_function(logfunc, NULL);
> >> +       libevdev_next_event(dev, LIBEVDEV_READ_NORMAL, NULL);
> >> +
> >> +       libevdev_set_log_function(logfunc, logdata);
> >
> > Shoudn't this be "libevdev_set_log_function(logfunc, NULL);"?
> > Otherwise, I don't get the "2" below.
> 
> You would get a NULL-deref in strcmp() in logfunc() if it's NULL. I
> think it looks all good. He tests _next_event() 3 times. First and
> last time with logfunc+logdata and in the middle without. So
> log_fn_called is 2 afterwards as the one in the middle is omitted.

yes, indeed. the point of this was to check that the logfunc is only called
when it is set, so  we set it, unset it, set it again and then expect that
it was called two times (not three times).

I've added a comment to that effect.

There should be another test checking for log data being valid, I'll add
that separately.

Cheers,
   Peter