oh! :D - the idea is simple: you don't know where a bug lies / if there's a bug at all - so you try out loads of random variations
-
when a test fails you print out the test to the console so it can be added & then fixed. E.g. https://github.com/yoshuawuyts/nanomorph/blob/a9236ceb4c2cafa11dcf593234cdee065bf01923/test.js#L242-L302 …
it gets even better if you don't use Math.random() but a seeded RNG instead; makes random tests reproducible :D
