[Nickle]Re: Nickle: array assignment?

[Keith Packard]

Around 21 o'clock on Feb 17, Barton C Massey wrote:
> Right now, after
>   a = [3]{1,2,3};
>   b = a;
> b is a reference to the same array as a.  Is this the
> "right" semantic?  If so, is there any way to make b be a
> copy of a, which is also often desirable?  In particular,
> if a is immutable, can I make b be a mutable copy of a?
> 
> I hate deep-copy/shallow-copy semantics.

I did shallow copy so that function arguments would work like C does.
Alternatively, we could always do deep-copy and pass references around 
instead.  But the additional syntax required is a pain:

	function bar (ref a)
	{
		(*a)[0] = 12;
	}

	array q = [3];
	bar(&q);
	

Oh, I know, I know!  We could use C++ semantics; I love them:

	foo (int &a)
	{
		a = 27;
	}

	bar ()
	{
		int	b;

		foo (b);
		return b;
	}
	
Those seem so clear and unambiguous to me.

Are there better boxing/unboxing syntax schemes?

keithp@keithp.com	 XFree86 Core Team		SuSE, Inc.