Expression switch - an alternate proposal

[forax at univ-mlv.fr]

----- Mail original -----
> De: "Cay Horstmann" <cay.horstmann at sjsu.edu>
> À: "Brian Goetz" <brian.goetz at oracle.com>, "Remi Forax" <forax at univ-mlv.fr>
> Cc: "amber-dev" <amber-dev at openjdk.java.net>
> Envoyé: Vendredi 13 Avril 2018 00:57:20
> Objet: Re: Expression switch - an alternate proposal

[...]

> 
> What about variable declarations? In statement switch, the following is
> legal:
> 
> switch (foo) {
>    case 1:
>       int x = 1;
>    default:
>       x = 2;
>       System.out.println(x);
> }
> 
> and the following is not:
> 
> switch (foo) {
>    case 1:
>       int x = 1;
>       System.out.println(x);
>       break;
>    case 2:
>       int x = 2;
>       System.out.println(x);
>       break;
> }
> 
> It seems that, if variable declarations are allowed in an expression
> switch, the rules are going to have to be different :-)
> 
> If you shore this up by restricting the statements that can follow a
> case label in an expression switch, haven't you implicitly defined a
> block expression? In that case, why not make it a general construct?


Here disabling fallthrough for the expression switch is an enabler.
Once there is no fallthrough, you can consider that the statements of a case are in an implicit block scope,
i.e
  var foo = switch(bar) {
    case 0:
      int x = 1;
      break x;
    default:
      break 42;
  };
to be equivalent to
  var foo = switch(bar) {
    case 0: {     // open a block scope
      int x = 1;
      break x;
    }             // end it
    default:
      break 42;
  };

but it can only works for an expression switch and if fallthroughs are disallowed, something we have not yet decided.

> 
> Cheers,
> 
> Cay
> 
> 

cheers,
Remi