Re: [PATCH] regex: use C99-style array arg syntax

[Bruno Haible]

Paul Eggert wrote:
> +/* Specify the number of elements of a function's array parameter,
> +   as in 'int f (int n, int a[__ARG_NELTS (n)]);'.  */
> +#if (defined __STDC_VERSION__ && 199901L <= __STDC_VERSION__ \
> +     && !defined __STDC_NO_VLA__)
> +# define __ARG_NELTS(n) n
> +#else
> +# define __ARG_NELTS(n)
> +#endif

The documentation of this macro makes it look like it expects an
expression as argument. But in the later uses you also often pass
the keyword 'static':

> +          regmatch_t pmatch[__ARG_NELTS (static nmatch)],
> +          regmatch_t prev_idx_match[__ARG_NELTS (static nmatch)],

How about changing the comment to

/* Specify the number of elements of a function's array parameter,
   as in 'int f (int n, int a[__ARG_NELTS (n)]);'
   or    'int f (int n, int a[__ARG_NELTS (static n)]);'  */


By the way, what is the point of adding the 'static' keyword? From [1][2][3]
I understand that when one writes 'TYPE a[static n]' it denotes
   a pointer to an array of at least n objects of type TYPE.
But since, for any m ≥ n, a pointer to m objects of type TYPE is also a
pointer to n objects of type TYPE, the words "at least" are redundant:
'TYPE a[static n]' denotes
   a pointer to an array of n objects of type TYPE.

So, what does 'TYPE a[n]' denote, then?
  - the same thing?
  - or a pointer to one object of type TYPE, just like 'TYPE a[]'?

Either way, the syntax 'TYPE a[n]' (without 'static') in function parameter
declarations appears redundant. Correct?

In that case, shouldn't the comment be changed to

/* Specify the number of elements of a function's array parameter, together
   with the 'static' keyword, as in
   'int f (int n, int a[__ARG_NELTS (static n)]);'  */

Bruno

[1] ISO C 11, § 6.7.6.3.(7)
[2] https://en.cppreference.com/w/c/language/array
[3] 
https://stackoverflow.com/questions/17559631/what-are-those-strange-array-sizes-and-static-in-c99