On Fri, Feb 7, 2025 at 2:33 AM David Gow davidgow@google.com wrote:
Thanks very much for doing this. I'm happy with these changes from a KUnit POV.
Two things I think we need to be careful about:
- This and the printf test are both changing the m68k configs. This is
fine, but could lead to a (harmless) merge conflict, so we should make that clear and try to avoid having them go up in separate trees. (And, if one gets merged first, rebase the other.)
- There has been some pushback on some kselftest->kunit conversions in
the past, especially if the test is being used to debug live systems (which typically don't have CONFIG_KUNIT enabled). I can't personally imagine that as an issue with scanf (though my imagination isn't perfect), so I'd doubt it's a problem.
I'm assuming that, as mentioned in v2, these will go in via printk, not ksefltest/kunit. Either would work for me (but, as mentioned above, I think this and the printf tests should go in via the same tree).
This series is: Reviewed-by: David Gow davidgow@google.com
Cheers, -- David
Thanks for the review David. Given the discussion on the printf series I applied the same scrutiny to this series; I reduced the churn, and kept the original control flow and failure messages.
I'll pick up your Reviewed-by and send v4 shortly.
Cheers. Tamir