@Linux the message I linked to leads with the most straightforward solution (which I'm not sure why you completely ignored), and I don't see a Discourse "edit" marker on it - although the person did make an "EDIT" (so maybe it was a "ninja edit") note with a perhaps-unfriendly comment.
I wouldn't liken a person doubting the premise of the question (with a very good reason - actually referring to the implementation which conflicts with what you're stating) to saying "your code is shit". Are you also yourself saying that your code is shit because you're asking for help with it, since it doesn't work as you expect? 🙂
I would say that "extraordinary claims require extraordinary evidence" here. If you're claiming that a program works in a wildly different way compared to:
* What the source code suggests
* What people have observed for years and years
it's on you to show it. NixOS, luckily, is a system where it should not be hard to make a self-contained, reproducible example.
I saw a similar thing in your thread about the `yes` utility, where you claimed (contrary to literally everyone else's experience) that a very well known program that has worked the same way for decades works very differently on different distros, and provided zero reproducible evidence.
It is pretty much impossible for others to prove a negative ("on no system in the world, at no point in time does this supposed bug happen"), but it should be trivially easy for you to prove a positive ("if you run this version of an OS, use this configuration, do this thing, then this happens") if you consistently encounter an issue.