On Fri, Apr 03, 2015 at 03:21:45PM +0200, Marek Chalupa wrote: > On Fri, Apr 3, 2015 at 4:16 AM, Bryce Harrington <[email protected]> > wrote: > > > config-malformed uses an invalid configuration file to verify that the > > test harness is indeed attempting to load the test's requested config > > file. > > + > > +#include "config.h" > > + > > +#include "weston-test-client-helper.h" > > + > > +TEST(config_malformed_test) > > +{ > > + /* We should never reach here, because the invalid config file > > + * should cause the server to exit with failure before invoking > > tests. > > + */ > > + assert(1 == 1); > > > > Maybe I don't understand what this test does, but assert(1 == 1) is always > true so this is kind of useless statement, isn't it? > Maybe you meant to use 1 != 1? And going further - couldn't be this fail > test converted to normal test, so that we are sure that the failure is > really > the one we test for?
Anything could be put there; the test itself never actually gets executed, because the failure occurs during the setup phase. I'm not sure if XFAIL is the right approach here, it's just the first thing I found that reliably worked. We also need a 'positive test', that is - set some parameter in the .ini file, then have the test verify the setting is being set. I'd love to do this with the mode parameter, but as I found this seems to not be recognized by the headless backend. A few other parameters I looked at could be set via .ini, but I didn't see how to query from the client-side. So, I leave all that for future work, but I think being able to verify set parameters is pretty important. Bryce _______________________________________________ wayland-devel mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/wayland-devel
