Currently it seems like just writing a test / expect for a definition suppresses the unused definition warning. Is that the intended behavior or is that likely to change? I personally think it should still warn on that and only stop on actual usage, not tests covering it.
I personally think it should still warn on that
Yeah, that makes sense, I'll make an issue for that. We can always close it if someone comes with a convincing opposing argument.
We already have an issue :)
#3238
I have thought about it as well. But sometimes you need a helper function for testing only
Last updated: Jul 26 2025 at 12:14 UTC