Date: Mon, 23 Jun 2014 17:15:03 -0700 From: Craig Rodrigues <rodrigc@FreeBSD.org> To: Warner Losh <imp@bsdimp.com> Cc: Brooks Davis <brooks@freebsd.org>, Dimitry Andric <dim@freebsd.org>, "Simon J. Gerraty" <sjg@juniper.net>, Glen Barber <gjb@freebsd.org>, freebsd-current Current <freebsd-current@freebsd.org>, Marcel Moolenaar <marcel@freebsd.org> Subject: Re: Problems building FreeBSD 9.2 on FreeBSD 10 Message-ID: <CAG=rPVcrJpP_1VcKZnsxAugBywJBXg63p2piX7nPndTunPEWDQ@mail.gmail.com> In-Reply-To: <8CD24B0A-DF45-4437-BEBE-8C67B241DE93@bsdimp.com> References: <CAG=rPVct-HzOksM2KQvPuOdf4zgZo__=Oannh0x%2B_JVHEzQyyw@mail.gmail.com> <CAG=rPVfkGa7=n-TSqcnP574ry0Au2THbqvAZJfKD5OK9a1FBNw@mail.gmail.com> <20140617200551.07542580A1@chaos.localdomain> <CAG=rPVeBgWS=vMGxSmX8O4QWW77XC2mqDgaiNrFsGrzg47UQYQ@mail.gmail.com> <20140617211436.GI1248@hub.FreeBSD.org> <CAG=rPVezMsPCP%2BB1ePw_Mh=Cv6qx07ordsWQ3cF4a_ytLWeCEQ@mail.gmail.com> <CAG=rPVd4evnkJoz1rVeMjP7YNKOW0BWF0MvSrFQ3u2xCiVEtmw@mail.gmail.com> <690CE378-D7D9-49A6-BC20-13FD540E63A2@FreeBSD.org> <CAG=rPVfanU=FVWPd768_G5NjXRHnUH5o05%2BKwFeLaTQqU3Ux8w@mail.gmail.com> <FC1B2983-4B49-4480-855E-6BD09B148F31@bsdimp.com> <CAG=rPVd5XGz8hfuBtcOuuZp8ND1ssWoTkVNn0Hg6Li_2-NrgcA@mail.gmail.com> <1ED3AC7E-0F74-46A7-BAAA-E30600DC23BB@bsdimp.com> <CAG=rPVeP32=y3VZgn0MXFPrr8tevL=jOsCow=rvpxRevmvoLqA@mail.gmail.com> <8CD24B0A-DF45-4437-BEBE-8C67B241DE93@bsdimp.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, OK, I think I see the issue. I looked here: http://svnweb.freebsd.org/base/stable/9/share/mk/bsd.prog.mk?view=3Dlog and saw that dim@ MFC'd his EARLY_BUILD stuff in r257812. That is why you can build stable/9 on a stable/10 host. I am building FreeBSD 9.2 which doesn't have that change. bsd.prog.mk in stable/9 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D .if defined(PROG_CXX) && !defined(EARLY_BUILD) .if !empty(CXXFLAGS:M-stdlib=3Dlibc++) echo ${PROG}: ${LIBCPLUSPLUS} >> ${DEPENDFILE} .else echo ${PROG}: ${LIBSTDCPLUSPLUS} >> ${DEPENDFILE} .endif .endif bsd.prog.mk in 9.2 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D .if defined(PROG_CXX) .if !empty(CXXFLAGS:M-stdlib=3Dlibc++) echo ${PROG}: ${LIBCPLUSPLUS} >> ${DEPENDFILE} .else echo ${PROG}: ${LIBSTDCPLUSPLUS} >> ${DEPENDFILE} .endif .endif bsd.prog.mk in CURRENT =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D .if defined(PROG_CXX) .if ${COMPILER_TYPE} =3D=3D "clang" && empty(CXXFLAGS:M-stdlib=3Dlibstdc++) echo ${PROG}: ${LIBCPLUSPLUS} >> ${DEPENDFILE} .else echo ${PROG}: ${LIBSTDCPLUSPLUS} >> ${DEPENDFILE} .endif .endif So, I guess that stable/9 can build properly on a stable/10 box. For FreeBSD 9.2, there is no easy way out. -- Craig On Mon, Jun 23, 2014 at 4:23 PM, Warner Losh <imp@bsdimp.com> wrote: > > On Jun 23, 2014, at 5:19 PM, Craig Rodrigues <rodrigc@FreeBSD.org> wrote: > > > > > > > > > On Mon, Jun 23, 2014 at 4:13 PM, Warner Losh <imp@bsdimp.com> wrote: > > > > Which bombing out are you seeing (two or three have been sighted in thi= s > thread)? And is this a nanobsd build, or a straight buildworld? > > > > When building FreeNAS, with a hacked the nanobsd > > script to does "make make buildworld", and the make.conf which I posted= , > I > > am seeing this: > > > > -------------------------------------------------------------- > > >>> stage 1.2: bootstrap tools > > -------------------------------------------------------------- > > cd /zroot/build/r/freenas2/FreeBSD/src; > MAKEOBJDIRPREFIX=3D/zroot/build/r/freenas2/os-base/amd64/zroot/build/r/fr= eenas2/FreeBSD/src/tmp > INSTALL=3D"sh /zroot/build/r/freenas2/FreeBSD/src/tools/install.sh" > PATH=3D/zroot/build/r/freenas2/os-base/amd64/zroot/build/r/freenas2/Free= BSD/src/tmp/legacy/usr/sbin:/zroot/build/r/freenas2/os-base/amd64/zroot/bui= ld/r/freenas2/FreeBSD/src/tmp/legacy/usr/bin:/zroot/build/r/freenas2/os-bas= e/amd64/zroot/build/r/freenas2/FreeBSD/src/tmp/legacy/usr/games:/zroot/buil= d/r/freenas2/os-base/amd64/zroot/build/r/freenas2/FreeBSD/src/tmp/legacy/bi= n:/sbin:/bin:/usr/sbin:/usr/bin > WORLDTMP=3D/zroot/build/r/freenas2/os-base/amd64/zroot/build/r/freenas2/= FreeBSD/src/tmp > VERSION=3D"9.3-ALPHA" MAKEFLAGS=3D"-m > /zroot/build/r/freenas2/FreeBSD/src/tools/build/mk -j 9 > .MAKE.LEVEL.ENV=3DMAKELEVEL NO_CLEAN=3D1 SRCCONF=3D/dev/null > __MAKE_CONF=3D/zroot/build/r/freenas2/os-base/amd64/make.conf.build -m > /zroot/build/r/freenas2/FreeBSD/src/share/mk TARGET=3Damd64 > TARGET_ARCH=3Damd64" COMPILER_TYPE=3Dclang > /zroot/build/r/freenas2/os-base/amd64/zroot/build/r/freenas2/FreeBSD/src/= make.amd64/make > -f Makefile.inc1 DESTDIR=3D BOOTSTRAPPING=3D1100022 SSP_CFLAGS=3D > -DWITHOUT_HTML -DWITHOUT_INFO -DNO_LINT -DWITHOUT_MAN -DNO_PIC > -DWITHOUT_PROFILE -DNO_SHARED -DNO_CPU_CFLAGS -DNO_WARNS -DNO_CTF > -DEARLY_BUILD bootstrap-tools > > =3D=3D=3D> gnu/usr.bin/gperf (obj,depend,all,install) > > =3D=3D=3D> gnu/usr.bin/gperf/doc (obj) > > =3D=3D=3D> gnu/usr.bin/gperf/doc (depend) > > make: don't know how to make /usr/lib/libstdc++.a. Stop > > *** [bootstrap-tools] Error code 2 > > 1 error > > *** [_bootstrap-tools] Error code 2 > > 1 error > > *** [buildworld] Error code 2 > > > > make[1]: stopped in /zroot/build/r/freenas2/FreeBSD/src > > 1 error > > > > make[1]: stopped in /zroot/build/r/freenas2/FreeBSD/src > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > > ERROR: build FAILED; see above or log file here: > /zroot/build/r/freenas2/os-base/amd64/_.bw > > *** Error code 2 > > > > Stop. > > make: stopped in /zroot/build/r/freenas2 > > > > > > > > > > > > And I specifically was doing my testing on -current, not 10.x. I haven'= t > back ported much of anything I've done to the build system, and if anybod= y > else has, then it is on them to make it work in the 10.x environment. Whi= le > it has usually worked, 9 on 10 isn't in the supported matrix we've > traditionally had in this project. > > > > I reproduced the same problem using a CURRENT build host ( 11.0-CURRENT > FreeBSD 11.0-CURRENT #1 r267305 ). > > > > I wonder how this could possibly happen on stable-10, since EARLY_BUILD i= s > still there to preclude the line being added. > > I'll have to re-run my test WITHOUT_CLANG. I just used the defaults. > > Any chance you can narrow the number of options required to trigger this? > > Warner > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAG=rPVcrJpP_1VcKZnsxAugBywJBXg63p2piX7nPndTunPEWDQ>