Ralf Wildenhues wrote:
> > What's the point of that?
>
> To supply a prototype for strcpy which is used in the test.
Thanks for explaining. The only risk is that of a possible prototype
clash for stpncpy() between <string.h> and the expected one, but I've
checked AIX and glibc (the only platforms known so far to have stpncpy())
and there appears to be no problem. So I checked in this:
*** stpncpy.m4 2 Mar 2005 17:19:45 -0000 1.6
--- stpncpy.m4 25 Sep 2006 14:33:50 -0000
***************
*** 1,5 ****
! # stpncpy.m4 serial 3
! dnl Copyright (C) 2002-2003, 2005 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
dnl with or without modifications, as long as this notice is preserved.
--- 1,5 ----
! # stpncpy.m4 serial 4
! dnl Copyright (C) 2002-2003, 2005-2006 Free Software Foundation, Inc.
dnl This file is free software; the Free Software Foundation
dnl gives unlimited permission to copy and/or distribute it,
dnl with or without modifications, as long as this notice is preserved.
***************
*** 22,27 ****
--- 22,29 ----
AC_CACHE_CHECK([for working stpncpy], gl_cv_func_stpncpy, [
AC_TRY_RUN([
#include <stdlib.h>
+ #include <string.h> /* for strcpy */
+ /* The stpncpy prototype is missing in <string.h> on AIX 4. */
extern char *stpncpy (char *dest, const char *src, size_t n);
int main () {
const char *src = "Hello";
