DMD 0.161 release

[Oskar Linde]

Walter Bright wrote:
> Jarrett Billingsley wrote:
>> "Walter Bright" <newshound at digitalmars.com> wrote in message 
>> news:e7832r$g4h$1 at digitaldaemon.com...
>>> Mostly bug fixes.
>>>
>>> http://www.digitalmars.com/d/changelog.html
>>
>> Oh man, those delegate literals look like it would be very possible to 
>> make interesting "pseudo-structures," that is, fake language 
>> constructs.  So if you were to define a function as
>>
>> void func(void delegate() dg);
>>
>> That is, the last parameter is a void delegate(), it would be an 
>> interesting bit of syntactic sugar to be able to write
>>
>> func
>> {
>>     writefln("foo");
>> }
>>
>> Not sure what utility this would present, but hey! 
> 
> There was some thought about doing that, but I'm not so sure it wouldn't 
> be more confusing than useful.

With the new delegate syntax, and a very simple function one can finally 
replace the annoying and common case where a for-loop until now still 
has to be used:

void repeat(Int,Delegate)(Int n, Delegate d) {
	static assert(is(Int:int),"Argument 1 to repeat must be int");
	for (int i = 0; i < n; i++)
		d();
}

The implementation could also contain:
static if (NumberOfArgs!(Delegate) == 1)
   d(i);

The use is:

repeat(10, { something(); });

And the multi-line version becomes:

repeat(10, {
   something();
   somethingElse();
});

Which at least makes me wish for Jarretts suggested short form:

repeat(10) {
   something();
   somethingElse();
}

But the implications of having such power is almost scary... :)

/Oskar