This would help with debugging. Each test outcome can report whether it was considered a pass/fail, its return code, and, if it was a failure, the reason for that failure (e.g., bad-return-code, bad-output, unable-to-decode). Note that the stdout/stderr won't be recorded in the outcome as that would consume vast resources for some of the scenarios in our dataset.
This would help with debugging. Each test outcome can report whether it was considered a
pass/fail, its return code, and, if it was a failure, the reason for that failure (e.g.,bad-return-code,bad-output,unable-to-decode). Note that the stdout/stderr won't be recorded in the outcome as that would consume vast resources for some of the scenarios in our dataset.Motivated by #145