2 ms·
This kind of thinking is sadly lost on many. I have seen copious amounts of nonsensical tests slapped full of hard-wired mocks and any change in the functionali
by janosdebugs 3y ago
This kind of thinking is sadly lost on many. I have seen copious amounts of nonsensical tests slapped full of hard-wired mocks and any change in the functionality would break hundreds of tests. In such a scenario an LLM might be the bandaid many are looking for. Then again, the value of such tests is questionable.
- robryk 3y agoMy best example of that was a test that was asserting that a monitoring metric changes in some way with a comment expressing doubt whether the asserted values are correct (but whoever wrote it still wrote the test that enshrined the behaviour that they themselves doubted).