How complex of a system can you test with property-based testing? More complex than you'd think (but not unlimited).
On using Google's American Fuzzy Lop for fuzz testing.
The theft Ruby gem allows for property-based testing — generating valid, randomly-generated input and validating that a particular property is true.
theft is a a C library for property-based testing: It generates input to stress-test code, and can reduce failures detected to minimal failing input.
Finding good test input can be tricky. Even with loads of unit tests, bugs still get through. Why not just generate input and let that find the edge cases?