Hi Shuah, On 25/06/2026 18:22, Matthieu Baerts (NGI0) wrote: > When there are many tests and/or many log lines, it is difficult to > quickly find out how the different tests went when running: > > make -C tools/testing/selftests run_tests > > This run_tests target executes 'run_many' from runner.sh, which relies > on ktap_helpers.sh. At the end of the execution, the KTAP_CNT_* counters > have been populated, and the total can then be simply printed with the > ktap_print_totals helper. > > Now a summary is printed at the end, e.g. here with the two last lines: > > $ make -sC tools/testing/selftests TARGETS=(...) run_tests > # timeout set to (...) > # selftests: (...) > (...) > ok 1 selftests: (...) > # timeout set to (...) > # selftests: (...) > (...) > ok 2 selftests: (...) # SKIP > # timeout set to (...) > # selftests: (...) > (...) > not ok 3 selftests: (...) > # 1 skipped test(s) detected. Consider enabling relevant config options > to improve coverage. > # Totals: pass:1 fail:1 xfail:0 xpass:0 skip:1 error:0 > > This summary is printed, using '#' as prefix to be treated as a comment > in the TAP format. The summary is also printed when summary=1 is used to > redirect log lines in different files because it still looks useful when > there are many tests, not to have to check each line. > > Note that this 'make' command will continue to exit without errors > (returned code is 0) when the tests can be executed. I guess it is > better not to change this behaviour to be able to continue detecting > kselftests failures. Gentle ping about this small addition? :)
Cheers, Matt -- Sponsored by the NGI0 Core fund.
