What if we can only keep the logs for failed tests? That way it would simplify downloading artifacts on the CRS test pipeline.
One way could be:
- run a test
- if it goes good, zero the error log file
- if there is a failure, concatenate it to a new file "failed-tests.log" for example
- repeat for every test
- in the end, there should be an empty error log, and maybe a file "failed-tests.log" if something failed.
- probably it would make sense only if the flag
--show-failures-only is enabled.
What if we can only keep the logs for failed tests? That way it would simplify downloading artifacts on the CRS test pipeline.
One way could be:
--show-failures-onlyis enabled.