On Wed, Aug 19, 2020 at 7:24 PM Brendan Higgins <brendanhiggins@xxxxxxxxxx> wrote: > > On Wed, Aug 19, 2020 at 12:39 PM Vitor Massaru Iha <vitor@xxxxxxxxxxx> wrote: > > > > In some cases, to maintain the consistency of the Expected messages > > with the original runtime test, it is necessary to customize the > > Expected messages on KUnit. > > > > As an example test_overflow conversion to KUnit (I added 1, > > p->s_of+1, just to fail the test). > > > > Using KUNIT_EXPECT_EQ: > > > > Expected _of == p->s_of+1, but > > _of == 0 > > p->s_of+1 == 1 > > not ok 1 - overflow_calculation_test > > ok 2 - overflow_shift_test > > > > Using KUNIT_EXPECT_EQ_CUSTOM_MSG: > > > > Expected 0 + 0 to not overflow (type u8) > > not ok 1 - overflow_calculation_test > > > > Which is more similar to the error message of the original test. > > > > Signed-off-by: Vitor Massaru Iha <vitor@xxxxxxxxxxx> > > All in all this looks pretty reasonable to me; however, is there any > chance you could provide an example of its usage? Sure I'll do that. I am reviewing the errors reported by 0day in the overflow_kunit patch v3, and I noticed when compiling overflow_kunit with the parameters of 0day and this patch, there were some errors related to this patch that I sent. I'll fix it too.
