> tests might fail one run and pass another. People don't seem to like this I have also encountered this attitude and haven't understood it so far. What stops

> tests might fail one run and pass another. People don't seem to like this I have also encountered this attitude and haven't understood it so far. What stops you from recording the test cases to reproduce failed ones?