[Ezbench-dev] [PATCH 11/25] test/piglit: do not report "completed" if the report generation failed
Petri Latvala
petri.latvala at intel.com
Fri Feb 24 11:19:12 UTC 2017
From: Martin Peres <martin.peres at linux.intel.com>
---
tests.d/piglit/piglit.test | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/tests.d/piglit/piglit.test b/tests.d/piglit/piglit.test
index 6e247c2..3e825ee 100644
--- a/tests.d/piglit/piglit.test
+++ b/tests.d/piglit/piglit.test
@@ -36,6 +36,8 @@ for test in tests:
print(test)
END
+ local exit_code=$?
+
# Go through all the subtests we wanted to get and check if all of them
# are present. If not, mark them as missing.
for subtest in "${run_sub_tests[@]}"; do
@@ -44,7 +46,7 @@ END
done
# Display the final status, if it was a full run
- if [ ! -f "${run_log_file}.testlist" ]; then
+ if [ "$exit_code" -eq 0 ] && [ ! -f "${run_log_file}.testlist" ]; then
echo ": completed"
fi
--
2.9.3
More information about the Ezbench-dev
mailing list