[Mono] Rename Fposmod to PosMod, fix output

[aaronfranke]

#18364 (comment)

1. None of the other float-returning methods start with the letter F

2. It's not PascalCase if you count pos and mod as words

3. The pos distinction is not necessary because for "non-positive modulus" also known as "remainder" there's already an operator in C# called %. People using the Mathf function would only ever be looking for something that wasn't in the C# language. In the interest of clarity and somewhat keeping consistency with the C++ math, I've kept the word Pos.

4. PosMod() looks a lot cleaner than Fposmod().

Also, I've changed the code inside the method. In my tests, this version works much better. The previous version of the code produces incorrect output for some values.

As someone who uses this function a lot in my code, I would definitely appreciate this change.

[neikeq]

There is nothing in the name Mod() that tells me it's any different from the modulo operator. I'm in favor of PosMod() or PositiveMod().

[aaronfranke]

% is not a Modulo operator. In C languages, % performs a Remainder operation. Explanation

Does it help to have inline documentation? Editors such as MonoDevelop can read this. We could add this for other methods later, if we want.

Also, I've realized it would be useful to have an integer version of this method, so I've added one.

[aaronfranke]

By the way, test code:

Console.WriteLine("Common use cases, where b is positive");
Console.WriteLine("4 Mod 5 = " + Mod(4, 5));
Console.WriteLine("5 Mod 5 = " + Mod(5, 5));
Console.WriteLine("6 Mod 5 = " + Mod(6, 5));
Console.WriteLine("-6 Mod 5 = " + Mod(-6, 5));
Console.WriteLine("-5 Mod 5 = " + Mod(-5, 5));
Console.WriteLine("-4 Mod 5 = " + Mod(-4, 5));
Console.WriteLine("Rare use cases, where b is negative");
Console.WriteLine("4 Mod -5 = " + Mod(4, -5));
Console.WriteLine("5 Mod -5 = " + Mod(5, -5));
Console.WriteLine("6 Mod -5 = " + Mod(6, -5));
Console.WriteLine("-6 Mod -5 = " + Mod(-6, -5));
Console.WriteLine("-5 Mod -5 = " + Mod(-5, -5));
Console.WriteLine("-4 Mod -5 = " + Mod(-4, -5));

Output:

Common use cases, where b is positive
4 Mod 5 = 4
5 Mod 5 = 0
6 Mod 5 = 1
-6 Mod 5 = 4
-5 Mod 5 = 0
-4 Mod 5 = 1
Rare use cases, where b is negative
4 Mod -5 = -1
5 Mod -5 = 0
6 Mod -5 = -4
-6 Mod -5 = -1
-5 Mod -5 = 0
-4 Mod -5 = -4

Output should be increasing, but looping around range ([0, b)), with increasing input (a).

With old Fposmod() code:

Common use cases, where b is positive
4 Mod 5 = 4
5 Mod 5 = 0
6 Mod 5 = 1
-6 Mod 5 = 4
-5 Mod 5 = 5  # Wrong
-4 Mod 5 = 1
Rare use cases, where b is negative
4 Mod -5 = 4  # Wrong
5 Mod -5 = 0  
6 Mod -5 = 1  # Wrong
-6 Mod -5 = -6  # Very Wrong
-5 Mod -5 = -5  # Wrong
-4 Mod -5 = -9  # Very Wrong

[KellyThomas]

The example from the docs indicates that it should return within the range (0, b].

http://docs.godotengine.org/en/3.0/classes/[email protected]#class-gdscript-fposmod

[aaronfranke]

@KellyThomas This is inconsistent with the positive values, and thus with the wording "wraps equally in positive and negative". I've submitted another PR to fix the C++ code: #19279

[aaronfranke]

@neikeq I've changed the function name to PosMod instead of Mod as you requested.

Since you've already concluded that the fix is correct, we should merge so that we have a consistent C# API before Godot 3.1 Alpha on July 1st. Unlike in the C++ code, nothing depends on this function (yet).

[neikeq]

Looks good to me. However, if this code is a mirror of the new implementation being introduced in #19279, then I would prefer both PRs to be merged at the same time to avoid problems.

[hpvb]

Cherry picked into 3.0.6