Check the return code for the test script.
Rework the error reporting from the dataset log analysis to take
into account global failures as well, and only report success if
both global success and dataset success have been achieved.
Fixes bug: 1263828
Change-Id: I53aa5ab949293a02f93f9981c0f98246a77d825f
diff --git a/turbo_hipster/lib/utils.py b/turbo_hipster/lib/utils.py
index 0555ca7..3739b44 100644
--- a/turbo_hipster/lib/utils.py
+++ b/turbo_hipster/lib/utils.py
@@ -184,6 +184,7 @@
logger.removeHandler(log_handler)
log_handler.flush()
log_handler.close()
+ return p.returncode
def push_file(job_log_dir, file_path, publish_config):