Re: [current tinderbox] failure on powerpc/powerpc

Peter Grehan wrote:
> >cc -O2 -pipe  -DIPFILTER_LOG -DIPFILTER_COMPILED -DIPFILTER_LOOKUP  
> >-DIPFILTER_SCAN -DIPFILTER_SYNC -DIPFILTER_CKSUM -I. 
> >-I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/tools -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter -DSTATETOP -D__UIO_EXPOSE -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized -c /tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/radix.c
> >cc -O2 -pipe  -DIPFILTER_LOG -DIPFILTER_COMPILED -DIPFILTER_LOOKUP  
> >-DIPFILTER_SCAN -DIPFILTER_SYNC -DIPFILTER_CKSUM -I. 
> >-I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/tools -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter -DSTATETOP -D__UIO_EXPOSE -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized -c /tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/bpf_filter.c
> >cc -O2 -pipe  -DIPFILTER_LOG -DIPFILTER_COMPILED -DIPFILTER_LOOKUP  
> >-DIPFILTER_SCAN -DIPFILTER_SYNC -DIPFILTER_CKSUM -I. 
> >-I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/tools -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../sys/contrib/ipfilter -DSTATETOP -D__UIO_EXPOSE -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized  -o ipftest ipftest.o fil.o ip_frag.o ip_state.o ip_nat.o ip_proxy.o ip_auth.o ip_htable.o ip_lookup.o ip_pool.o ip_scan.o ip_sync.o ip_rules.o ip_fil.o ip_log.o ippool_y.o ippool_l.o ipf_y.o ipf_l.o ipnat_y.o ipnat_l.o md5.o radix.o bpf_filter.o -L/tinderbox/CURRENT/powerpc/powerpc/obj/powerpc/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../libipf -lipf -lkvm
> >gzip -cn 
> >/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipftest/../../../contrib/ipfilter/man/ipftest.1 > ipftest.1.gz
> >===> sbin/ipf/ipmon (all)
> >cc -O2 -pipe  -DLOGFAC=LOG_LOCAL0 -I. 
> >-I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../contrib/ipfilter -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../contrib/ipfilter/tools -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../sys -I/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../sys/contrib/ipfilter -DSTATETOP -D__UIO_EXPOSE -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized -c /tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../contrib/ipfilter/tools/ipmon.c
> >/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../contrib/ipfilter/tools/ipmon.c: In function `print_ipflog':
> >/tinderbox/CURRENT/powerpc/powerpc/src/sbin/ipf/ipmon/../../../contrib/ipfilter/tools/ipmon.c:1030: warning: comparison is always false due to limited range of data type
> 
>  I've sent mail to Darren about this: it's a PPC-only (well, maybe ARM 
> as well) default-unsigned-char problem. For those that want to 
> buildworld appended is the patch I'm using.
> 
> later,
> 
> Peter.
> 
> RCS file: /home/ncvs/src/contrib/ipfilter/tools/ipmon.c,v
> retrieving revision 1.2
> diff -u -r1.2 ipmon.c
> --- ipmon.c     25 Apr 2005 18:20:15 -0000      1.2
> +++ ipmon.c     1 May 2005 08:42:53 -0000
> _at__at_ -1024,7 +1024,7 _at__at_
>         (void) sprintf(t, "%*.*s%u", len, len, ipf->fl_ifname, 
> ipf->fl_unit);
>         t += strlen(t);
>  #endif
> -#ifdef __sgi
> +#if (defined(__sgi) || defined(__powerpc__))
>         if ((ipf->fl_group[0] == 255) && (ipf->fl_group[1] == '\0'))
>  #else
>         if ((ipf->fl_group[0] == -1) && (ipf->fl_group[1] == '\0'))

While I suppose it might be possible to drop the #if completely, it
would be better to use #if CHAR_MIN == 0 here.  That would make it work
for FreeBSD/arm too, as well as any platform where chars are unsigned by
default on which this code may run.

Cheers,
Maxime