RCS1089: += 1 is not always equivalent to postfix increment

[jhinder]

public class C {
    int i;
    
    public int M()
    {
        i += 1; // <-- equivalent to i++
        return i += 1; // <-- equivalent to ++i!
    }
}

Sharplab

Whenever the value of the += expression is taken, it's equivalent to a pre-increment, not a post-increment. When applying the code fix RCS1089, this difference can introduce subtle off-by-one bugs.

Version: 2.3.0

[josefpihrt]

It's the same with yield return.

SharpLab

Do you know of any other syntax where the pre-increment should be used?

[jhinder]

Passing arguments (M(i += 1)), array indexing (a[i += 1]) -- basically anything where the += expression is used as an expression and not just a statement, by the looks of it.

And of course it's not just the increment operator, this affects the decrement operator as well.

[josefpihrt]

I came to this conclusion:

If an assignment expression is used as a part of statement, post-increment should be used:
i += 1; > i++;

Every other use of i += 1 should be replaced with pre-increment.

Same logic applies to decrement operation.