No, because I may end up building a kernel that has no INET6 for a cut down platform that has no src.conf option set. -a On 7 September 2014 19:34, Glen Barber <gjb_at_freebsd.org> wrote: > On Sun, Sep 07, 2014 at 03:34:35PM -0700, yaneurabeya_at_gmail.com wrote: >> On Sep 7, 2014, at 14:08, Glen Barber <gjb_at_FreeBSD.org> wrote: >> >> > On Sun, Sep 07, 2014 at 02:03:35PM -0700, Sean Bruno wrote: >> >> make[4]: "/home/sbruno/bsd/fbsd_head/sys/modules/if_gif/Makefile" line >> >> 12: warning: Couldn't read shell's output for >> >> "cat /home/sbruno/bsd/obj/mips/mips.mips/home/sbruno/bsd/fbsd_head/sys/WZR-300HP/opt_inet6.h" >> >> >> >> >> >> Shouldn't this cat be done in a saner way? >> >> >> > >> > This is done quite often throughout the tree, and bmake(1) does not like >> > it when there is no value assigned from the '!=' expansion. >> > >> > The fix I've seen most commonly done is to echo a newline after the >> > assignment, such as: >> > >> > Index: sys/modules/if_gif/Makefile >> > =================================================================== >> > --- sys/modules/if_gif/Makefile (revision 271215) >> > +++ sys/modules/if_gif/Makefile (working copy) >> > _at__at_ -9,7 +9,7 _at__at_ KMOD= if_gif >> > SRCS= if_gif.c in_gif.c opt_inet.h opt_inet6.h opt_mrouting.h >> > >> > .if defined(KERNBUILDDIR) >> > -OPT_INET6!= cat ${KERNBUILDDIR}/opt_inet6.h >> > +OPT_INET6!= cat ${KERNBUILDDIR}/opt_inet6.h; echo >> > .if empty(OPT_INET6) >> > MK_INET6_SUPPORT=no >> > .endif >> >> Shouldn’t this all be removed and replaced with equivalent logic provided by kern.opts.mk? > > Probably. > > Glen >Received on Mon Sep 08 2014 - 00:56:25 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:52 UTC