Re: standards/45258: #include <string.h> -> <strings.h> -> <stdint.h>

[Klaus Klein <kleink%kleink.org@localhost>]

On Tue, Aug 16, 2011 at 01:40:04AM +0000, David Holland wrote:
> The following reply was made to PR standards/45258; it has been noted by 
> GNATS.
> 
> From: David Holland <dholland-bugs%netbsd.org@localhost>
> To: gnats-bugs%NetBSD.org@localhost
> Cc: 
> Subject: Re: standards/45258: #include <string.h> -> <strings.h> -> <stdint.h>
> Date: Tue, 16 Aug 2011 01:39:10 +0000
> 
>  On Tue, Aug 16, 2011 at 01:30:00AM +0000, 
> dholland%eecs.harvard.edu@localhost wrote:
>   > Either remove <strings.h> from <string.h>, which is probably not a
>   > viable proposition, or fix <strings.h> not to spew into the namespace.
>  
>  This should do it, but I haven't tested for build fallout yet. There
>  will probably be some.
>  
>  Any objections?

If that (fixing <strings.h>) does the trick, certainly not from me.
In that case I'd also suggest something like this (similar to the
printf(3) approach to varargs):

Index: popcount.3
===================================================================
RCS file: /cvsroot/src/lib/libc/string/popcount.3,v
retrieving revision 1.4
diff -u -r1.4 popcount.3
--- popcount.3  22 Jul 2009 07:07:27 -0000      1.4
+++ popcount.3  16 Aug 2011 07:31:01 -0000
@@ -27,7 +27,7 @@
 .\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 .\" POSSIBILITY OF SUCH DAMAGE.
 .\"
-.Dd July 13, 2009
+.Dd August 16, 2011
 .Dt POPCOUNT 3
 .Os
 .Sh NAME
@@ -47,6 +47,7 @@
 .Fn popcountl "unsigned long value"
 .Ft unsigned int
 .Fn popcountll "unsigned long long value"
+.In stdint.h
 .Ft unsigned int
 .Fn popcount32 "uint32_t value"
 .Ft unsigned int


- Klaus