Thomas Huth <thuth@xxxxxxxxxx> writes:
> On 13/11/2019 16.06, Alex Bennée wrote:
>>
>> Thomas Huth <thuth@xxxxxxxxxx> writes:
>>
>>> While working on the travis.yml file, I've run into cases where
>>> all tests are reported as "SKIP" by the run_test.sh script (e.g.
>>> when QEMU could not be started). This should not result in a
>>> successful test run, so mark it as failed if not at least one
>>> test passed.
>>
>> But doesn't this mean you could have everything fail except one pass and
>> still report success?
>
> The FAILs are already handled one line earlier...
Oops I should have noticed that:
Reviewed-by: Alex Bennée <alex.bennee@xxxxxxxxxx>
>
>>>
>>> Signed-off-by: Thomas Huth <thuth@xxxxxxxxxx>
>>> ---
>>> .travis.yml | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/.travis.yml b/.travis.yml
>>> index 9ceb04d..aacf7d2 100644
>>> --- a/.travis.yml
>>> +++ b/.travis.yml
>>> @@ -115,3 +115,4 @@ script:
>>> - make -j3
>>> - ACCEL="${ACCEL:-tcg}" ./run_tests.sh -v $TESTS | tee results.txt
>>> - if grep -q FAIL results.txt ; then exit 1 ; fi
>
> ... here -----------^
>
>>> + - if ! grep -q PASS results.txt ; then exit 1 ; fi
>
> Maybe it would also be nicer to provide proper exit values in the
> run_tests.sh script, but the logic there gives me a bad headache...
> grep'ing for FAIL and PASS in the yml script is way more easy.
>
> Thomas
--
Alex Bennée
