Port tslint-microsoft-contrib's no-increment-decrement

[JoshuaKGoldberg]

See #1142.

I remember reading a book (JavaScript: The Good Parts? Clean Code? The Art of Unit Testing?) by one of the big influential programming folks and encountering a passage arguing in favor of this type of rule. The justification was something like these points:

• It's easy to type +i or -i instead of --i
• Standardizing widths of increments/decrements makes them easier to read
• It's easy to forget the difference between ++i and i++

Perhaps this rule should be named increment-decrement and only ban the pre- operators, to offend fewer people? The first point is the only one of the three I feel would withstand a real debate.

[ajafff]

It's easy to type +i or -i instead of --i

when used as an ExpressionStatement, no-unused-expression would warn you about that.
This is only an issue when assigning the result.
But even then you get an error from prefer-const because you never reassign the variable (if this is the only place you might have modified it).

[JoshuaKGoldberg]

Counter-example:

const items = ["apple", "banana", "cherry"];
let index = 0;

while (+index) {
    // ^
    if (items[index][0] === "b") {
        ++index;
    }
}

[JoshuaKGoldberg]

Fixed by #3575.