Modular increment/decrement operators

[Rabadash8820]

I often find myself writing code that has to traverse a list and "loop around". E.g., in game code, I might want to pick 5 random entities from a list of N to initiate some behavior; a naive/performant implementation is to pick 5 elements starting from a random index in the list, with modulus to avoid IndexOutOfBoundsException. Having a single operator to do this "increment/decrement index but loop around" would be quite handy.

Here is the idea for the "modular increment operator":

int index = <random #>;

// Current C# operators:
index = (index + 1) % count;
// or
index++;
index %= count;

// Suggested operator
index ++%= count;

A "modular decrement operator", --%= would work similarly. Its implementation may not be as simple as index = (index - 1) % count as it would have to account for negative numbers (I typically use index = (index - 1 + count) % count in these situations). I think it could be implemented as index = (index - 1) %% count if I understand the proposal in #4744 correctly.

I admit, a 4-character operator is kinda ridiculous, and incrementing/decrementing by a number other than 1 would still require the same code as before. I did consider a ternary assignment operator like index +%= delta count but I'm not sure that's any more readable than current C# options. Anyway, just wanted to see if anyone else has considered this idea or would find it worthwhile.

[CyrusNajmabadi]

could do something like:

    public static void IncrementAndWrap(this ref int index, int count)
    {
        index = (index + 1) % count;
    }

[Rabadash8820]

True, thanks. Always cool to see a ref use case. It would be nice to avoid the method-call overhead for something this small, but some MathUtil static class could have a method like this. I'd probably just stick with inlining over this though. 🤷‍♂️

[AartBluestoke]

probably no overhead at all- the compiler trivially in-lines this.

[colejohnson66]

Indeed, trivially inlined, even at tier 0. Don't take the JIT (or NativeAOT) for granted; It's as powerful as your normal C compiler when it comes to analysis.

[Rabadash8820]

Wo, never seen sharplab before. Thanks for the link @colejohnson66, and for verifying this!