Variable-length arrays

[Peter S. Shenkin]

From: Stephen D. Clamage <uunet!taumet.com!steveauunet.UU.NET>
[[ Comments deleted concerning problems which may arise with 
	float f( double a[n], int n );
These are the first such comments I've seen that indicate that this syntax 
might cause problems for the user, as distinguished from the compiler-writer.  
Perhaps someone from Cray would care to reply?  However, I see possible 
problems with Stephen's modest proposal, which was:  ]]

> I have a modest proposal.  Follow the example of ISO Pascal conformant
> array parameters.  Declare the size parameter of the array inside the
> brackets:
> 	float f(double a[int n]);

How is this function to be called?  Do you mean to imply that it should be
called as follows:
	var = f( array[43] );  /* within f, n is assigned value 43 */
If so, how would I pass an array to f whose first element is array[12], and
which has a total of 43 elements, starting with array[12]?  If I understand
you correctly, I would have to say:
	var = f( (&(array[12]))[43] );
I don't have the precedence rules in my head; thus probably I have more parens
in this expression than are necessary;  nevertheless, this seems awkward and 
non-intuitive for the programmer.  Aside from the awkwardness, it also 
requires a specific calling order, which I think most of us feel is 
undesirable.  If we're going to require a specific calling order, I would
much rather revert to the simpler form:
	float f( int n, double a[n] );
where n has to be declared first.

But I would like to see someone from Cray respond to Stephen's remark about
difficulties in the simple prototype "float f( double a[n], int n )" if
there should be a typedef around called n.

	-P.
	-P.
************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb*************************
Peter S. Shenkin, Department of Chemistry, Barnard College, New York, NY 10027
(212)854-1418    shenkinaavogadro.barnard.columbia.edu   shenkinacunixc.BITNET
********** "I've got algorithm -- who could ask for anything more?" **********