On Wed, Jan 16, 2013 at 11:44:22PM +0100, Dimitry Andric wrote: > On 2013-01-11 00:41, Brooks Davis wrote: > > On Sun, Dec 23, 2012 at 01:54:08AM +0100, Dimitry Andric wrote: > >> On 2012-12-18 17:37, Brooks Davis wrote: > >>> Author: brooks > >>> Date: Tue Dec 18 16:37:24 2012 > >>> New Revision: 244401 > >>> URL: http://svnweb.freebsd.org/changeset/base/244401 > >>> > >>> Log: > >>> Replace our implementation of the vis(3) and unvis(3) APIs with > >>> NetBSD's. This output size limited versions of vis and unvis > >>> functions > >>> as well as a set of vis variants that allow arbitrary characters to be > >>> specified for encoding. > >> > >> This seems to break bootstrapping in some scenarios, in "stage 4.2: > >> building libraries"; for example, with a test run with gcc on > >> ref10-amd64.f.o I got this: > >> > >> gcc -O2 -pipe -I/scratch2/tmp/dim/head/lib/libc/include > >> -I/scratch2/tmp/dim/head/lib/libc/../../include > >> -I/scratch2/tmp/dim/head/lib/libc/amd64 -DNLS -D__DBINTERFACE_PRIVATE > >> -I/scratch2/tmp/dim/head/lib/libc/../../contrib/gdtoa -DINET6 > >> -I/scratch2/tmp/dim/obj/scratch2/tmp/dim/head/lib/libc > >> -I/scratch2/tmp/dim/head/lib/libc/resolv -D_ACL_PRIVATE -DPOSIX_MISTAKE > >> -I/scratch2/tmp/dim/head/lib/libc/../../contrib/jemalloc/include > >> -I/scratch2/tmp/dim/head/lib/libc/../../contrib/tzcode/stdtime > >> -I/scratch2/tmp/dim/head/lib/libc/stdtime > >> -I/scratch2/tmp/dim/head/lib/libc/locale -DBROKEN_DES -DPORTMAP > >> -DDES_BUILTIN -I/scratch2/tmp/dim/head/lib/libc/rpc -DYP -DNS_CACHING > >> -D_FREEFALL_CONFIG -DSYMBOL_VERSIONING -std=gnu99 -fstack-protector > >> -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized > >> -Wno-pointer-sign -c > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/vis.c -o vis.o > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c: In > >> function 'unvis': > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:237: error: > >> 'VIS_NOESCAPE' undeclared (first use in this function) > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:237: error: > >> (Each undeclared identifier is reported only once > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:237: error: > >> for each function it appears in.) > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:241: error: > >> 'VIS_HTTP1808' undeclared (first use in this function) > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:245: error: > >> 'VIS_HTTP1866' undeclared (first use in this function) > >> /scratch2/tmp/dim/head/lib/libc/../../contrib/libc-vis/unvis.c:249: error: > >> 'VIS_MIMESTYLE' undeclared (first use in this function) > >> > >> There should most likely be an explicit -I option to point the compiler > >> at the correct vis.h header during the early stages, otherwise it will > >> pick up /usr/include/vis.h, which does not have several of these new > >> VIS_XXX defines. > > > > Sorry for not following up on this sooner. I've not seen this at all > > and most of my builds are done on an 9.0-STABLE box so vis.h isn't > > updated. Adding > > > > CFLAGS+= -I${CURDIR}/../../contrib/libc-vis > > > > to lib/libc/gen/Makefile.inc seems like it should fix this, but I'm > > surprised not to have bumped into this or broken tinderbox. Is there > > anything odd about your buildworld command? > > Not really, the environment was just this: > > 1) Building on ref10-amd64.f.o > 2) CC=gcc, CXX=g++, CPP=gcpp > > Did you manage to reproduce the failure?
I never did with a buildworld. I could if I tried to build by hand after touching unvis.c so I added an entry to CFLAGS. I think it's more correct. -- Brooks
pgpcOQXZ8iYih.pgp
Description: PGP signature
