Closed Wilfred closed 8 years ago
Hmm. The failures should print the arguments with which the test ran. That should be sufficient to reproduce the failure?
You're right, that's totally sufficient. Sorry for the noise.
For more complex test-input, it is not always that simple to refer to the debug output. It would be very useful if quickcheck could present reproducability steps on a failing test-case.
In the best of all worlds, giving a single one-liner that runs just the failing test-case with the failing test-input would be awesome.
I have tests that fail very rarely. When this happens on Travis, I'd really like to reproduce the issue locally.
If quickcheck randomly chose a seed, but printed that seed value on failure, then I could replicate test failures.