Removing mandatory break from switch cases

[osamakawish]

I'm not even sure why it's there... but I have examples of code that could be a whole lot cleaner if I could skip the break statement.

Any reason for why it's mandatory?

[HaloFour]

Yes, the break is required to avoid the common bug of accidentally falling through cases which occurs in C and Java. If you want fallthrough you can use goto to jump to a specific case label.

[osamakawish]

Nvm... I think I misunderstood the original switch...

[HaloFour]

Nvm... I think I misunderstood the original switch...

No worries, I think that's exactly why the original C# designers made the break mandatory. They wanted to keep with the C syntax and behavior but prevent occurrences of fallthrough where the developer didn't intend for it to happen. Looking back they definitely could have made the break implicit and achieved the same thing.

[CyrusNajmabadi]

they definitely could have made the break implicit

Yes. but then what would this mean:

switch (x)
{
case 0:
case 1:
   ...

if 'break' were implicit, then case 0 woudln't fallthrough to 1. If we wanted that, we'd effectively be saying: break is implicit except for the empty case. And i think that would have been more confusing. How we have it today is how i prefer it. You break when you break. You fallthrough when tehre's no break. Fallthrough is only allowed in the trivial case.

[HaloFour]

Yep, plus the argument that behaving differently than Java or C would have been confusing.

[HaloFour]

I also imagine that if the team had considered removing implicit fallthrough that they would have also considered different syntax for that scenario, such as case 0, 1:

[HaloFour]

Switch fallthrough never reevaluates the condition. The switch jumps to the first case based on the expression and then executes all following statements until it hits a break or other control statement to leave the switch block.

Example in Java:

  public class JavaFiddle
  {
    public static void main(String[] args)
    {
        int i = 1;
        
        switch (i) {
            case 1: System.out.println("Hello");
            case 2: System.out.println("World");
            default: System.out.println("Fallthrough");
        }
    }
  }

prints:

Hello
World
Fallthrough

Example in C:

#include <stdio.h>

int main()
{
    int i = 1;
    
    switch (i) {
        case 1: printf("Hello\n");
        case 2: printf("World\n");
        default: printf("Fallthrough\n");
    }
}

prints:

Hello
World
Fallthrough

[canton7]

The confusion might be related to the fact that case statements which use patterns or when clauses are tested in order

[IS4Code]

It remains to specify whether omitting the break means fallthrough or implicit break. I have nothing against implicit break, but it may cause confusion when compared with other C-like languages. It might also be inconsistent with multiple consecutive case labels, as this is one of the occasions where fallthrough is allowed.

[wes-sleeman]

It might be nice (and hopefully not lead to confusion) to allow implicit breaks only for the last case. If that were downgraded from an error to a warning (to be honest, I wouldn't mind it being removed entirely, but that's the VB programmer in me), I don't think people would be too upset and it would certainly save those of us coming from other languages a bit of a headache if we didn't care. Can't speak for others, but needing the break at the end has always irked me.

[qrli]

In the early days, most C# programmers come from C/C++/Java, and a lot of code are ported from C/C++/Java. The strictness of switch helped to prevent common mistakes. But now, most new C# programmers have no C experience, and the switch legacy feels alien.

I wish switch expression to replace switch statement in 99% use cases. New comers don't need to know quirks of switch statement. E.g. its fall-though behavior, as it is rarely used, even in legacy code.