Re: [PATCH] lib: kunit: add test_min_heap test conversion to KUnit

[Ian Rogers]

On Mon, Oct 12, 2020 at 2:03 PM Brendan Higgins
<brendanhiggins@xxxxxxxxxx> wrote:
>
> On Tue, Aug 4, 2020 at 9:22 AM Vitor Massaru Iha <vitor@xxxxxxxxxxx> wrote:
> >
> > Hi Peter,
> >
> > On Tue, Aug 4, 2020 at 11:23 AM <peterz@xxxxxxxxxxxxx> wrote:
> > >
> > > On Tue, Aug 04, 2020 at 10:46:21AM -0300, Vitor Massaru Iha wrote:
> > > > On Tue, Aug 4, 2020 at 10:25 AM <peterz@xxxxxxxxxxxxx> wrote:
> > > > > On Wed, Jul 29, 2020 at 06:57:17PM -0300, Vitor Massaru Iha wrote:
> > > > >
> > > > > > The results can be seen this way:
> > > > > >
> > > > > > This is an excerpt from the test.log with the result in TAP format:
> > > > > > [snip]
> > > > > > ok 5 - example
> > > > > >     # Subtest: min-heap
> > > > > >     1..6
> > > > > >     ok 1 - test_heapify_all_true
> > > > > >     ok 2 - test_heapify_all_false
> > > > > >     ok 3 - test_heap_push_true
> > > > > >     ok 4 - test_heap_push_false
> > > > > >     ok 5 - test_heap_pop_push_true
> > > > > >     ok 6 - test_heap_pop_push_false
> > > > > > [snip]
> > >
> > > So ^ is TAP format?
> >
> > Yep, you can see the spec here: https://testanything.org/tap-specification.html
> >
> > >
> > > > > I don't care or care to use either; what does dmesg do? It used to be
> > > > > that just building the self-tests was sufficient and any error would
> > > > > show in dmesg when you boot the machine.
> > > > >
> > > > > But if I now have to use some damn tool, this is a regression.
> > > >
> > > > If you don't want to, you don't need to use the kunit-tool. If you
> > > > compile the tests as builtin and run the Kernel on your machine
> > > > the test result will be shown in dmesg in TAP format.
> > >
> > > That's seems a lot more verbose than it is now. I've recently even done
> > > a bunch of tests that don't print anything on success, dmesg is clutter
> > > enough already.
> >
> > What tests do you refer to?
> >
> > Running the test_min_heap.c, I got this from dmesg:
> >
> > min_heap_test: test passed
> >
> > And running min_heap_kunit.c:
> >
> > ok 1 - min-heap
>
> Gentle poke. I think Vitor was looking for a response. My guess is
> that Ian was waiting for a follow up patch.

There were some issues in the original patch, they should be easy to
fix. I'm more concerned that Peter's issues are addressed about the
general direction of the patch, verbosity and testing frameworks. I
see Vitor followed up with Peter but I'm not sure that means the
approach has been accepted.

Thanks,
Ian