Earlier quoted context omitted.
Not mutation testing — nothing is mutated. It builds both branches with instrumentation, runs the same test suite on each, and diffs the recorded arguments and return values. In that example, selectDiscount returns a different discount and the order total goes 85 → 60. Neither function is in the diff; only the sort helper is. The tests still pass — one asserts the total is positive, and 60 is positive.
please try speaking in your own words. i still have no idea what this is for. is it two compare two impls? is it to "optimize" a function (i assume not since there's no mutation)?
hi, this is simply put, to compare the behavior diff of two builds. when i say behavior, i mean the return values of functions etc. it'll help with AI-assisted code reviews etc. since AI looks at the text diff only, and misses behavior diff downstream. it complements static analysis and unit tests for a stronger testing strategy.