placeInAppropriate0To360Scope is a lot more complicated than it needs to be. I simplified it greatly (which also improves its readability) and added a more formal definition to the documentation for clarification.
Aside from the fact that the expected logic is the same as the original function, I performed 100 billion fuzz tests with random angles in a 64-revolution range and my new function does not deviate from the original function except for minuscule differences (less than 10^-12) due to floating point error that arises because the new implementation performs fewer operations, as well as in scenarios with multiple correct outputs (e.g. the original function returns 0 when provided 180.0, 0.0, while mine returns 360).
This was also tested in my school's 2024 competition (well I just graduated from there but you get the point) and it worked perfectly there.
thenetworkgrinch
commented
4 months ago
placeInAppropriate0To360Scopeis a lot more complicated than it needs to be. I simplified it greatly (which also improves its readability) and added a more formal definition to the documentation for clarification.Aside from the fact that the expected logic is the same as the original function, I performed 100 billion fuzz tests with random angles in a 64-revolution range and my new function does not deviate from the original function except for minuscule differences (less than 10^-12) due to floating point error that arises because the new implementation performs fewer operations, as well as in scenarios with multiple correct outputs (e.g. the original function returns
0when provided180.0, 0.0, while mine returns360).This was also tested in my school's 2024 competition (well I just graduated from there but you get the point) and it worked perfectly there.