Alternative syntax for closures

[Rémi Forax]

Neal Gafter a écrit :
> One idea we tossed around a few months ago, without any real 
> conclusion, was to replace the curly braces with parens. Things have 
> changed a bit since then; I'd now use a dash-greater instead of 
> equals-greater.  I think we could consider using the parens for 
> restricted closures (for those of you who haven't been paying 
> attention, restricted closures can't use return, break, or continue to 
> a target outside the closure) and curly braces for the unrestricted form.
>
> /RestrictedLambdaExpression/
> ( /FormalParameterDecls////_/opt/ /// *—**>* ////BlockStatements_/opt 
> / ////////Expression///////_/opt/ /////// ////)
> /RestrictedFunctionType/
> ( /TypeList////_/opt/ /// *—**>* /Type/ /ThrowsClause////_/opt/ /// )
>
> /UnrestrictedLambdaExpression/
> { /FormalParameterDecls////_/opt/ /// *—**>* ////BlockStatements_/opt 
> / ////////Expression///////_/opt/ /////// ////}
> /UnrestrictedFunctionType/
> { /TypeList////_/opt/ /// *—**>* /Type/ /ThrowsClause////_/opt/ /// }
>
> For example (this is /not/ intended to be a motivational example):
>
> *(int, int **—**> int) plus = (int x, int y **—**> x+y);*
> *{ Answer —**> void } block = { **Answer answer —**> 
> System.out.println(answer); };*
>
> The parens instead of curly braces emphasize to the reader that it is 
> an expression closure, which is one of the principal use cases for the 
> restricted form.
>
> Without an active working group, I don't expect to make any major 
> syntax changes in the prototype, but I wanted to get this out there 
> for you to think about.
>
> Regards,
> Neal

I've just written a blog about an alternative syntax:
http://weblogs.java.net/blog/forax/archive/2008/07/alternative_syn.html

Remi