Jakub Kicinski <kuba@xxxxxxxxxx> writes: > On Fri, 17 Jan 2025 13:49:23 +0100 Petr Machata wrote: >> This logging and xfail handling duplicates lib.sh. Would a patch like >> below be OK with you? The gist of it is to just use check_err, log_test >> and xfail_on_slow from lib.sh to achieve what the test open-codes. > > Hm, maybe? I'm not going to say no if you send a patch, but IMHO > if we were to change the output I think we should make it ktap. Hmm, at some point, we'll want to convert lib.sh to ktap as well. Hopefully transparently so that all the existing selftests out there just magically become ktap-compatible. If you want to preserve the output 1:1 then I feel like that's bending over backwards too much. We could still maybe reuse xfail_on_slow and ask for FAIL_TO_XFAIL, but that's marked internal in lib.sh. So I retract my proposal. Too faffy at that point.
