[Nickle] About arrays and things

[Keith Packard]

Around 1 o'clock on Jun 30, Bart Massey wrote:

>    It probably
>    depends on whether one views the i in the typedef to be
>    bound at type-def time, or at type-use time.  Nickle
>    chooses the latter view.

Isn't this backwards?  I think nickle should choose to evaluate all 
expressions where they occur and not where the value is used.

         int i = 3;
         typedef int[i] t;
	 t a1 = {1,2,3};

This should be invalid -- typedefs will (for at least a while) be 
evaluated in global scope, so 'i' will not be available.  Instead:

	global int i = 3;
	typedef int[i] t;
	t a1 = { 1, 2, 3 };
	i = 4;
	t a2 = { 1, 2, 3 };	/* t is still an array of 3 ints */

Currently, the dimension expression is compiled whereever a declaration 
uses it.  This is clearly wrong -- it will be evaluated as many times as 
it is used:

	global int i = 3;
	typedef int [++i] t;

	for (int j = 0; j = 7; j++) {
	    t q;
	    bar (q);
	}

If we use 'late' evaluation (as nickle currently does), we get 'bar' 
called with ever larger arrays...

-keith