On Mon, Jan 02, 2006 at 15:31:16 +0100, Martin Husemann wrote:

> I'm not sure this causes any problems, but just in case:
> While looking at some strange C compiler warnings about duplicate defintions
> I came along  this output from the src/tools/compat configure:
> 
> configure:7660: checking for htole16
> configure:7717: cc -o conftest -O   conftest.c -lz  >&5
> /var/tmp//ccYmqDa7.o(.text+0x8): In function `main':
> : undefined reference to `htole16'
> /var/tmp//ccYmqDa7.o(.text+0xc): In function `main':
> : undefined reference to `htole16'
> /var/tmp//ccYmqDa7.o(.data+0x0): undefined reference to `htole16'
> 
> and it does not define HAVE_HTOLE16 then.
> 
> This is on sparc64, -current as of a few hours ago, and of course there is
> a htole16() macro in <sys/endian.h> (or <machine/endian.h>).
> 
> Should this be fixed?

This is a very, very old problem.  ISTR, wrong autoconf macros used to
test for htole* &co, but last time I looked I decided I don't have
enough of autoconf fu to fix it properly and it doesn't really bother
me that much.

Since we provide re-implementation anyway, the easiest way is to alway
use out own version, I'd guess.

SY, Uwe
-- 
uwe@ptc.spbu.ru                         |       Zu Grunde kommen
http://snark.ptc.spbu.ru/~uwe/          |       Ist zu Grunde gehen