lazy redux

[Denis Koroskin]

On Mon, 07 Dec 2009 16:51:29 +0300, Michal Minich  
<michal.minich at gmail.com> wrote:

> Hello Denis,
>
>>> 1. auto t = new Thread ( { a=42 } );
>>> or auto t = new Thread ( () { a=42 } );
>> It already works, just try it (but don't forget to put a semicolon at
>> the  end).
>>
>>> 2. array.findAll (arr, (item) { item.contains ("abc") } ); 3. foo (
>>> (a, b) { a + b } );
>>>  4. array.findAll (arr, (item) { return item.contains ("abc"); } );
>>>  I'm not proposing this syntax (maybe I probably should, but I have
>>> feeling I would not be first). It may not even be possible to parse
>>> it,  but seems to me more similar to how currently functions are
>>> written. In  this setting {exp} or {stm} is not *special* case.
>>>
>> I believe it would work. And yes, it was already proposed by many
>> others.
>>
>
> it works with two differences:
>
> 1. the semicolon is required, even if the body consist only of one  
> expression. This is a minor detail.
> 2. more importantly - parameter types must be specified explicitly. I  
> don't understand why type of b cannot be inferred in this example:
>
> void foo (void delegate (int a) dg)
> {
>     dg(1);
> }
>
> void main ()
> {
>     foo ( (int b)  { writeln (b); } );
> }
>
> Does the type inference has problem with template code or some other  
> combination of features...?
>
>

It complicates semantic pass quite significantly: you can't semantically  
analyze delegate until you analyze "foo". It also can't infer arguments  
types in some cases:

void foo(T)(void delegate(T) dg);
foo( (a) { a.bar(); } );

Although I believe it is implementable and worth the trouble, there is a  
little gain in this feature and that's probably why it is low in the list.  
I think that Walter will give a green light if someone implements the  
feature and provides a complete patch.

Any volunteers?