Re: ports/181913: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous
On Sat, 7 Sep 2013 22:49:54 GMT rak...@freebsd.org wrote: Synopsis: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous State-Changed-From-To: open-patched State-Changed-By: rakuco State-Changed-When: Sat Sep 7 22:47:43 UTC 2013 State-Changed-Why: I don't think the previous version worked. From your description, it looks like you've switched to building with libc++ whereas libstdc++ was being used before. The upcoming Qt 4.8.5 plus a few patches which only made it to 4.8.6 (but we've backported) will finally make Qt build with libc++. We've just sent an exp-run request for Qt 4.8.5, and will hopefully fix all these errors once it is committed. http://www.freebsd.org/cgi/query-pr.cgi?pr=181913 I build the world/kernel since early this year with CXXFLAGS+= -stdlib=libc++ CXXFLAGS+= -std=c++11 in /etc/src.conf. I do not use those flags in /etc/make.conf! /etc/src.conf is supposed to target ONLY the /usr/src world, not the ports - this is as I interpret the man page for /etc/src.conf and it would be logical. But this rule/thinking seems to be broken by some includes from /usr/ports/Mk ingredients. I can assure that I didn't switch anything to build the ports but rebuilding world and then restarting building. Something must have changed since then in the logic of how libc++ slipped in instead of of libstdc++. What I did was a make delete-old-files, which deleted several GNU gcc stuff on all CURRENT boxes. I did not see that any lib got killed after I tried make delete-old-libs. And I did not check whether libstdc++ is still being built. There are many other occasions where now c++ errors occur and I guess those ports need to be reported in one by one via PR? signature.asc Description: PGP signature
net/openldap24-server: libtool: compile: gcc -g -O2 -Wall -DDO_SAMBA -I../../../inc [...] eval: gcc: not found *** Error code 1
After having built a world on CURRENT r255356 and having done make delete-old the outdated gnuish compiler stuff gets deleted. After that, port net/openldap[-sasl]-server rejects to compile due to the shown error. Somehwere the compiler seems to be hardcoded. [...] --- all-common --- for page in slapacl.8slapadd.8 slapauth.8 slapcat.8 slapd.8 slapdn.8 slapindex.8 slappasswd.8 slapschema.8slaptest.8; do sed -e s%LDVERSION%2.4.36% -e 's%ETCDIR%/usr/local/etc/openldap%g' -e 's%LOCALSTATEDIR%/var/db%' -e 's%SYSCONFDIR%/usr/local/etc/openldap%' -e 's%DATADIR%/usr/local/share/openldap%' -e 's%SBINDIR%/usr/local/sbin%' -e 's%BINDIR%/usr/local/bin%' -e 's%LIBDIR%/usr/local/lib%' -e 's%LIBEXECDIR%/usr/local/libexec%' -e 's%MODULEDIR%/usr/local/libexec/openldap%' -e 's%RELEASEDATE%2013/08/17%' ./$page | (cd .; soelim -) $page.tmp; done /usr/local/bin/libtool --mode=compile gcc -g -O2 -Wall -DDO_SAMBA -I../../../include -I../../../include -I../../../servers/slapd -I/usr/heimdal/include -c smbk5pwd.c libtool: compile: gcc -g -O2 -Wall -DDO_SAMBA -I../../../include -I../../../include -I../../../servers/slapd -I/usr/heimdal/include -c smbk5pwd.c -fPIC -DPIC -o .libs/smbk5pwd.o eval: gcc: not found *** Error code 1 Stop. make[2]: stopped in /usr/ports/net/openldap24-server/work/openldap-2.4.36/contrib/slapd-modules/smbk5pwd *** Error code 1 Stop. make[1]: stopped in /usr/ports/net/openldap24-server *** Error code 1 Stop. make: stopped in /usr/ports/net/openldap24-server === make failed for net/openldap24-server === Aborting update === Killing background jobs signature.asc Description: PGP signature
Re: ports/181913: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous
08.09.2013 10:14, O. Hartmann пишет: I can assure that I didn't switch anything to build the ports but rebuilding world and then restarting building. Something must have changed since then in the logic of how libc++ slipped in instead of of libstdc++. What I did was a make delete-old-files, which deleted several GNU gcc stuff on all CURRENT boxes. I did not see that any lib got killed after I tried make delete-old-libs. And I did not check whether libstdc++ is still being built. Yes, recently gcc was switched off by default at CURRENT, so libc++ is used both for the system and ports. There are many other occasions where now c++ errors occur and I guess those ports need to be reported in one by one via PR? Guess so. Preferrable with a patch. ;-) -- WBR, Boris Samorodov (bsam) FreeBSD Committer, http://www.FreeBSD.org The Power To Serve ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Re: bsd.port.pre.mk vs bsd.port.options.mk
On Sat, Sep 7, 2013 at 2:42 PM, Christian Weisgerber na...@mips.inka.dewrote: I have port that does something like .include bsd.port.pre.mk .if ${ARCH} == ... ... .endif .include bsd.port.post.mk A while back somebody submitted a PR asking me to replace bsd.port.pre.mk with bsd.port.options.mk, because it also makes ARCH available and is far less expensive. Now, a priori it is not clear to me that including options.mk is actually cheaper than pre.mk. And it seems odd to include options.mk but then not use any part of the options framework. The Porter's Handbook explicitly mentions ARCH as one of the variables provided by pre.mk. What's the preferred way to handle this? -- Christian naddy Weisgerber na...@mips.inka.de It is preferred to evaluate ARCH with bsd.port.options.mk. -jgh -- Jason Helfman | FreeBSD Committer j...@freebsd.org | http://people.freebsd.org/~jgh | The Power to Serve ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Troubles with dependencies (file -L is broken?)
Hi, I am trying to portupgrade my subversion and it keeps trying to install databases/db42 over already installed version (same version), missing the shared library dependency. Adding some debug into bsd.port.mk, I see: === subversion-1.8.3 depends on package: libtool=2.4 - found set -x; set -e ; for i in libdb-4.2.so:/usr/ports/databases/db42; do lib=${i%%:*} ; dir=${i#*:} ; target=install; depends_args=; echo -n === subversion-1.8.3 depends on shared library: ${lib} ; found=0 ; dirs=/lib /usr/lib /usr/local/lib `/bin/cat /usr/local/libdata/ldconfig/* 2/dev/null || : ` ; for libdir in $dirs; do test -f ${libdir}/${lib} || continue; if [ -x /usr/bin/file ]; then [ `file -b -L --mime-type ${libdir}/${lib}` = application/x-sharedlib ] || continue ; fi ; found=1 ; echo - found; done ; if [ ${found} -eq 0 ]; then echo - not found; echo ===Verifying for $lib in $dir; if [ ! -d $dir ] ; then echo = No directory for $lib. Skipping..; else if [ -n -o -n ]; then subpkgfile=`(cd $dir; make $depends_args -V PKGFILE)`; subpkgname=${subpkgfile%-*} ; subpkgname=${subpkgname##*/} ; if [ -r ${subpkgfile} -a $target = install ]; then echo === Installing existing package ${subpkgfile}; if [ -n -a ${subpkgname} = pkg ]; then [ -d /usr/ports/devel/subversion/work ] || /bin/mkdir -p /usr/ports/devel/subversion/work ; /usr/bin/tar xf ${subpkgfile} -C /usr/ports/devel/subversion/work -s ,/.*/,,g */pkg-static ; /usr/ports/devel/subversion/work/pkg-static add ${subpkgfile}; /bin/rm -f /usr/ports/devel/subversion/work/pkg-static; else /usr/sbin/pkg_add ${subpkgfile}; fi; elif [ -n -a ${target} = install ]; then echo === subversion-1.8.3 depends on package: ${subpkgfile} - not found; echo === USE_PACKAGE_DEPENDS_ONLY set - will not build from source; exit 1; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; echo === Returning to build of subversion-1.8.3; fi ; fi ; done + set -e + lib=libdb-4.2.so + dir=/usr/ports/databases/db42 + target=install + depends_args='' + echo -n '=== subversion-1.8.3 depends on shared library: libdb-4.2.so' === subversion-1.8.3 depends on shared library: libdb-4.2.so+ found=0 + /bin/cat /usr/local/libdata/ldconfig/compat7x /usr/local/libdata/ldconfig/mysql /usr/local/libdata/ldconfig/portupgrade /usr/local/libdata/ldconfig/pth + dirs='/lib /usr/lib /usr/local/lib /usr/local/lib/compat /usr/local/lib/mysql /usr/local/lib/compat/pkg /usr/local/lib/pth' + test -f /lib/libdb-4.2.so + continue + test -f /usr/lib/libdb-4.2.so + continue + test -f /usr/local/lib/libdb-4.2.so + [ -x /usr/bin/file ] + file -b -L --mime-type /usr/local/lib/libdb-4.2.so + [ inode/symlink = application/x-sharedlib ] + continue + test -f /usr/local/lib/compat/libdb-4.2.so + continue + test -f /usr/local/lib/mysql/libdb-4.2.so + continue + test -f /usr/local/lib/compat/pkg/libdb-4.2.so + continue + test -f /usr/local/lib/pth/libdb-4.2.so + continue + [ 0 -eq 0 ] + echo ' - not found' - not found + echo '===Verifying for libdb-4.2.so in /usr/ports/databases/db42' ===Verifying for libdb-4.2.so in /usr/ports/databases/db42 So, file(1) call is the culprit here. Trying to reproduce in the console: [sobomax@pioneer ~]$ file -b -L --mime-type /usr/local/lib/libdb-4.2.so inode/symlink However: [sobomax@pioneer ~]$ hexdump -C /usr/local/lib/libdb-4.2.so | head -n 20 7f 45 4c 46 02 01 01 09 00 00 00 00 00 00 00 00 |.ELF| 0010 03 00 3e 00 01 00 00 00 d0 40 02 00 00 00 00 00 |...п@..| 0020 40 00 00 00 00 00 00 00 b8 a6 0e 00 00 00 00 00 |@...╦і..| 0030 00 00 00 00 40 00 38 00 05 00 40 00 1d 00 1a 00 |@.8...@ .| 0040 01 00 00 00 05 00 00 00 00 00 00 00 00 00 00 00 || 0050 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || 0060 dc 45 0e 00 00 00 00 00 dc 45 0e 00 00 00 00 00 |эE..эE..| 0070 00 00 20 00 00 00 00 00 01 00 00 00 06 00 00 00 |.. .| 0080 00 50 0e 00 00 00 00 00 00 50 2e 00 00 00 00 00 |.P...P..| 0090 00 50 2e 00 00 00 00 00 fc 36 00 00 00 00 00 00 |.P..Э6..| 00a0 d8 3a 00 00 00 00 00 00 00 00 20 00 00 00 00 00 |ь: .| 00b0 02 00 00 00 06 00 00 00 e8 54 0e 00 00 00 00 00 |ХT..| 00c0 e8 54 2e 00 00 00 00 00 e8 54 2e 00 00 00 00 00 |ХT..ХT..| 00d0 a0 01 00 00 00 00 00 00 a0 01 00 00 00 00 00 00 |═...═...| 00e0 08 00 00 00 00 00 00 00 50 e5 74 64 04 00 00 00 |PЕtd| 00f0 5c 39 0d 00 00 00 00 00 5c 39 0d 00 00 00 00 00 |\9..\9..| 0100 5c 39 0d 00 00 00 00 00 b4 34 00 00 00 00 00 00 |\9..Є4..| 0110 b4 34 00 00 00 00 00 00 04 00 00 00 00 00 00 00 |Є4..| The culprit is that /usr/local/lib/libdb-4.2.so is a symlink to another symlink to another symlink etc, so that my guess is
Re: Troubles with dependencies (file -L is broken?)
P.S. This is fresh 9.2-RC3 with /usr on ZFS. -Maxim On Sun, Sep 8, 2013 at 12:09 AM, Maxim Sobolev sobo...@freebsd.org wrote: Hi, I am trying to portupgrade my subversion and it keeps trying to install databases/db42 over already installed version (same version), missing the shared library dependency. Adding some debug into bsd.port.mk, I see: === subversion-1.8.3 depends on package: libtool=2.4 - found set -x; set -e ; for i in libdb-4.2.so:/usr/ports/databases/db42; do lib=${i%%:*} ; dir=${i#*:} ; target=install; depends_args=; echo -n === subversion-1.8.3 depends on shared library: ${lib} ; found=0 ; dirs=/lib /usr/lib /usr/local/lib `/bin/cat /usr/local/libdata/ldconfig/* 2/dev/null || : ` ; for libdir in $dirs; do test -f ${libdir}/${lib} || continue; if [ -x /usr/bin/file ]; then [ `file -b -L --mime-type ${libdir}/${lib}` = application/x-sharedlib ] || continue ; fi ; found=1 ; echo - found; done ; if [ ${found} -eq 0 ]; then echo - not found; echo ===Verifying for $lib in $dir; if [ ! -d $dir ] ; then echo = No directory for $lib. Skipping..; else if [ -n -o -n ]; then subpkgfile=`(cd $dir; make $depends_args -V PKGFILE)`; subpkgname=${subpkgfile%-*} ; subpkgname=${subpkgname##*/} ; if [ -r ${subpkgfile} -a $target = install ]; then echo === Installing existing package ${subpkgfile}; if [ -n -a ${subpkgname} = pkg ]; then [ -d /usr/ports/devel/subversion/work ] || /bin/mkdir -p /usr/ports/devel/subversion/work ; /usr/bin/tar xf ${subpkgfile} -C /usr/ports/devel/subversion/work -s ,/.*/,,g */pkg-static ; /usr/ports/devel/subversion/work/pkg-static add ${subpkgfile}; /bin/rm -f /usr/ports/devel/subversion/work/pkg-static; else /usr/sbin/pkg_add ${subpkgfile}; fi; elif [ -n -a ${target} = install ]; then echo === subversion-1.8.3 depends on package: ${subpkgfile} - not found; echo === USE_PACKAGE_DEPENDS_ONLY set - will not build from source; exit 1; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; echo === Returning to build of subversion-1.8.3; fi ; fi ; done + set -e + lib=libdb-4.2.so + dir=/usr/ports/databases/db42 + target=install + depends_args='' + echo -n '=== subversion-1.8.3 depends on shared library: libdb-4.2.so ' === subversion-1.8.3 depends on shared library: libdb-4.2.so+ found=0 + /bin/cat /usr/local/libdata/ldconfig/compat7x /usr/local/libdata/ldconfig/mysql /usr/local/libdata/ldconfig/portupgrade /usr/local/libdata/ldconfig/pth + dirs='/lib /usr/lib /usr/local/lib /usr/local/lib/compat /usr/local/lib/mysql /usr/local/lib/compat/pkg /usr/local/lib/pth' + test -f /lib/libdb-4.2.so + continue + test -f /usr/lib/libdb-4.2.so + continue + test -f /usr/local/lib/libdb-4.2.so + [ -x /usr/bin/file ] + file -b -L --mime-type /usr/local/lib/libdb-4.2.so + [ inode/symlink = application/x-sharedlib ] + continue + test -f /usr/local/lib/compat/libdb-4.2.so + continue + test -f /usr/local/lib/mysql/libdb-4.2.so + continue + test -f /usr/local/lib/compat/pkg/libdb-4.2.so + continue + test -f /usr/local/lib/pth/libdb-4.2.so + continue + [ 0 -eq 0 ] + echo ' - not found' - not found + echo '===Verifying for libdb-4.2.so in /usr/ports/databases/db42' ===Verifying for libdb-4.2.so in /usr/ports/databases/db42 So, file(1) call is the culprit here. Trying to reproduce in the console: [sobomax@pioneer ~]$ file -b -L --mime-type /usr/local/lib/libdb-4.2.so inode/symlink However: [sobomax@pioneer ~]$ hexdump -C /usr/local/lib/libdb-4.2.so | head -n 20 7f 45 4c 46 02 01 01 09 00 00 00 00 00 00 00 00 |.ELF| 0010 03 00 3e 00 01 00 00 00 d0 40 02 00 00 00 00 00 |...п@..| 0020 40 00 00 00 00 00 00 00 b8 a6 0e 00 00 00 00 00 |@...╦і..| 0030 00 00 00 00 40 00 38 00 05 00 40 00 1d 00 1a 00 |@.8...@ .| 0040 01 00 00 00 05 00 00 00 00 00 00 00 00 00 00 00 || 0050 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || 0060 dc 45 0e 00 00 00 00 00 dc 45 0e 00 00 00 00 00 |эE..эE..| 0070 00 00 20 00 00 00 00 00 01 00 00 00 06 00 00 00 |.. .| 0080 00 50 0e 00 00 00 00 00 00 50 2e 00 00 00 00 00 |.P...P..| 0090 00 50 2e 00 00 00 00 00 fc 36 00 00 00 00 00 00 |.P..Э6..| 00a0 d8 3a 00 00 00 00 00 00 00 00 20 00 00 00 00 00 |ь: .| 00b0 02 00 00 00 06 00 00 00 e8 54 0e 00 00 00 00 00 |ХT..| 00c0 e8 54 2e 00 00 00 00 00 e8 54 2e 00 00 00 00 00 |ХT..ХT..| 00d0 a0 01 00 00 00 00 00 00 a0 01 00 00 00 00 00 00 |═...═...| 00e0 08 00 00 00 00 00 00 00 50 e5 74 64 04 00 00 00 |PЕtd| 00f0 5c 39 0d 00 00 00 00 00 5c 39 0d 00 00 00 00 00 |\9..\9..| 0100 5c 39 0d 00 00 00 00 00 b4 34 00 00
Re: AFFECTS: 10-CURRENT users with any port depending on converters/libiconv
On 09/07/13 21:46, O. Hartmann wrote: On Sat, 07 Sep 2013 14:27:48 +0200 Guido Falsi madpi...@freebsd.org wrote: On 09/07/13 14:10, olli hauer wrote: There are 13 ports using --with-iconv=${LOCALBASE} devel/apr1 devel/apr2 devel/git irc/epic5 lang/gauche net-mgmt/ettercap net/ssltunnel-client net/yaz net/zebra-server textproc/libxml2 textproc/py-libxml2 www/apache22 www/apache24 Most of these do work anyway. I'm sure about various of these. I have them working on my PCs. This does not mean they don't need to be tweaked anyway, but they have lower priority. net-mgmt/ettercap is known broken and I have it in my pipe. I'm giving this all the time I can, but I can''t spend too much time on this right now. I'm going to check these and fix the broken ones asap. and devel/glib20, print/ghostscript8, print/ghostscript9 using --with-libiconv=gnu --with-libiconv=native --with-libiconv=no --with-libiconv=no glib20 I already fixed in the big commit. uses native or gnu where appropriate. I'll also have a look at the ghostscript ports asap, but at least ghostscript9 I have seen it working on my PCs. Unfortunately Uses/iconv.mk defines only --with-libiconv(-prefix). If Uses/iconv.mk can be extended with something like ICON_PATH, then the 13 ports can be changed quickly to use the right iconv. Most of those will use the right iconv anyway if only one is found. Extending iconv.mk should be discussed with portmgr, adding a variable shouldn't be a problem though. This happens in editors/abiword: libtool: link: ( cd .libs rm -f libimp.la ln -s ../libimp.la libimp.la ) gmake[7]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument/imp' gmake[6]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument/imp' gmake[6]: Entering directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' ../../doltlibtool --tag=CXX --mode=link c++ -O2 -pipe -O3 -march=native -fno-strict-aliasing -lgsf-1 -lgobject-2.0 -lglib-2.0 -lintl -L/usr/local/lib -lxml2 -lgthread-2.0 -pthread -lgobject-2.0 -L/usr/local/lib -lglib-2.0 -lintl -L../../src -labiword-2.8 -lz -avoid-version -module -no-undefined -L/usr/local/lib -o opendocument.la -rpath /usr/local/lib/abiword-2.8/plugins common/libcommon.la exp/libexp.la imp/libimp.la -ljpeg grep: /usr/local/lib/libiconv.la: No such file or directory sed: /usr/local/lib/libiconv.la: No such file or directory libtool: link: `/usr/local/lib/libiconv.la' is not a valid libtool archive gmake[6]: *** [opendocument.la] Error 1 gmake[6]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' gmake[5]: *** [all-recursive] Error 1 gmake[5]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' gmake[4]: *** [all-recursive] Error 1 gmake[4]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins' gmake[3]: *** [all-recursive] Error 1 gmake[3]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6' gmake[2]: *** [all] Error 2 gmake[2]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6' === Compilation failed unexpectedly. Try to set MAKE_JOBS_UNSAFE=yes and rebuild before reporting the failure to the maintainer. *** Error code 1 This one looks like you still have some old libtoool archive which references libiconv laying around. Can you try again this command line(I rewrite here for convenience): find /usr/local/lib -name '*.la' -exec grep -qi iconv {} \; -print | xargs -n 1 pkg which -oq | sort -u and force rebuild of any port which still shows up? If none shows up (which would be strange, but I can't exclude anything) Hope is not lost and there are still some things we can try. -- Guido Falsi madpi...@freebsd.org ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Re: AFFECTS: 10-CURRENT users with any port depending on converters/libiconv
On Sun, 08 Sep 2013 10:44:01 +0200 Guido Falsi madpi...@freebsd.org wrote: On 09/07/13 21:46, O. Hartmann wrote: On Sat, 07 Sep 2013 14:27:48 +0200 Guido Falsi madpi...@freebsd.org wrote: On 09/07/13 14:10, olli hauer wrote: There are 13 ports using --with-iconv=${LOCALBASE} devel/apr1 devel/apr2 devel/git irc/epic5 lang/gauche net-mgmt/ettercap net/ssltunnel-client net/yaz net/zebra-server textproc/libxml2 textproc/py-libxml2 www/apache22 www/apache24 Most of these do work anyway. I'm sure about various of these. I have them working on my PCs. This does not mean they don't need to be tweaked anyway, but they have lower priority. net-mgmt/ettercap is known broken and I have it in my pipe. I'm giving this all the time I can, but I can''t spend too much time on this right now. I'm going to check these and fix the broken ones asap. and devel/glib20, print/ghostscript8, print/ghostscript9 using --with-libiconv=gnu --with-libiconv=native --with-libiconv=no --with-libiconv=no glib20 I already fixed in the big commit. uses native or gnu where appropriate. I'll also have a look at the ghostscript ports asap, but at least ghostscript9 I have seen it working on my PCs. Unfortunately Uses/iconv.mk defines only --with-libiconv(-prefix). If Uses/iconv.mk can be extended with something like ICON_PATH, then the 13 ports can be changed quickly to use the right iconv. Most of those will use the right iconv anyway if only one is found. Extending iconv.mk should be discussed with portmgr, adding a variable shouldn't be a problem though. This happens in editors/abiword: libtool: link: ( cd .libs rm -f libimp.la ln -s ../libimp.la libimp.la ) gmake[7]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument/imp' gmake[6]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument/imp' gmake[6]: Entering directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' ../../doltlibtool --tag=CXX --mode=link c++ -O2 -pipe -O3 -march=native -fno-strict-aliasing -lgsf-1 -lgobject-2.0 -lglib-2.0 -lintl -L/usr/local/lib -lxml2 -lgthread-2.0 -pthread -lgobject-2.0 -L/usr/local/lib -lglib-2.0 -lintl -L../../src -labiword-2.8 -lz -avoid-version -module -no-undefined -L/usr/local/lib -o opendocument.la -rpath /usr/local/lib/abiword-2.8/plugins common/libcommon.la exp/libexp.la imp/libimp.la -ljpeg grep: /usr/local/lib/libiconv.la: No such file or directory sed: /usr/local/lib/libiconv.la: No such file or directory libtool: link: `/usr/local/lib/libiconv.la' is not a valid libtool archive gmake[6]: *** [opendocument.la] Error 1 gmake[6]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' gmake[5]: *** [all-recursive] Error 1 gmake[5]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins/opendocument' gmake[4]: *** [all-recursive] Error 1 gmake[4]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6/plugins' gmake[3]: *** [all-recursive] Error 1 gmake[3]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6' gmake[2]: *** [all] Error 2 gmake[2]: Leaving directory `/usr/ports/editors/abiword/work/abiword-2.8.6' === Compilation failed unexpectedly. Try to set MAKE_JOBS_UNSAFE=yes and rebuild before reporting the failure to the maintainer. *** Error code 1 This one looks like you still have some old libtoool archive which references libiconv laying around. Can you try again this command line(I rewrite here for convenience): find /usr/local/lib -name '*.la' -exec grep -qi iconv {} \; -print | xargs -n 1 pkg which -oq | sort -u and force rebuild of any port which still shows up? If none shows up (which would be strange, but I can't exclude anything) Hope is not lost and there are still some things we can try. On one box, the command sequence (thanks for being redundant, I didn't find the previous email with the sequence within) reveals: converters/psiconv editors/abiword math/gnumeric Deleteing psiconv and gnumeric makes the sequence not showing both ports again, but then, installing gnumeric again, which reels in psiconv, the output of the command sequence is a s shown. Something is strange here ... Another prerequisite port with oldish iconv reliances? signature.asc Description: PGP signature
libiconv.a(iconv.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object
Hi, (Please Cc: me when replying, I'm not subscribed.) I created a fresh FreeBSD 9.1-RELEASE amd64 jail. No installed ports. I cannot build devel/gettext. Does anyone experience the same problem? I invariably run into the following problem: libtool: link: cc -shared -fPIC -DPIC .libs/bindtextdom.o .libs/dcgettext.o .libs/dgettext.o .libs/gettext.o .libs/finddomain.o .libs/hash-string.o .libs/loadmsgcat.o .libs/localealias.o .libs/textdomain.o .libs/l10nflist.o .libs/explodename.o .libs/dcigettext.o .libs/dcngettext.o .libs/dngettext.o .libs/ngettext.o .libs/plural.o .libs/plural-exp.o .libs/localcharset.o .libs/threadlib.o .libs/lock.o .libs/relocatable.o .libs/langprefs.o .libs/localename.o .libs/log.o .libs/printf.o .libs/setlocale.o .libs/version.o .libs/xsize.o .libs/osdep.o .libs/intl-compat.o -L/usr/local/lib /usr/local/lib/libiconv.a -O2 -Wl,-soname -Wl,libintl.so.9 -o .libs/libintl.so.9 /usr/bin/ld: /usr/local/lib/libiconv.a(iconv.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/libiconv.a: could not read symbols: Bad value Full typescript here: http://people.freebsd.org/~jlh/typescript.gettext.txt -- Jeremie Le Hen Scientists say the world is made up of Protons, Neutrons and Electrons. They forgot to mention Morons. ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Re: Troubles with dependencies (file -L is broken?)
On Sun, Sep 08, 2013 at 12:09:32AM -0700, Maxim Sobolev wrote: Hi, I am trying to portupgrade my subversion and it keeps trying to install databases/db42 over already installed version (same version), missing the shared library dependency. Adding some debug into bsd.port.mk, I see: === subversion-1.8.3 depends on package: libtool=2.4 - found set -x; set -e ; for i in libdb-4.2.so:/usr/ports/databases/db42; do lib=${i%%:*} ; dir=${i#*:} ; target=install; depends_args=; echo -n === subversion-1.8.3 depends on shared library: ${lib} ; found=0 ; dirs=/lib /usr/lib /usr/local/lib `/bin/cat /usr/local/libdata/ldconfig/* 2/dev/null || : ` ; for libdir in $dirs; do test -f ${libdir}/${lib} || continue; if [ -x /usr/bin/file ]; then [ `file -b -L --mime-type ${libdir}/${lib}` = application/x-sharedlib ] || continue ; fi ; found=1 ; echo - found; done ; if [ ${found} -eq 0 ]; then echo - not found; echo ===Verifying for $lib in $dir; if [ ! -d $dir ] ; then echo = No directory for $lib. Skipping..; else if [ -n -o -n ]; then subpkgfile=`(cd $dir; make $depends_args -V PKGFILE)`; subpkgname=${subpkgfile%-*} ; subpkgname=${subpkgname##*/} ; if [ -r ${subpkgfile} -a $target = install ]; then echo === Installing existing package ${subpkgfile}; if [ -n -a ${subpkgname} = pkg ]; then [ -d /usr/ports/devel/subversion/work ] || /bin/mkdir -p /usr/ports/devel/subversion/work ; /usr/bin/tar xf ${subpkgfile} -C /usr/ports/devel/subversion/work -s ,/.*/,,g */pkg-static ; /usr/ports/devel/subversion/work/pkg-static add ${subpkgfile}; /bin/rm -f /usr/ports/devel/subversion/work/pkg-static; else /usr/sbin/pkg_add ${subpkgfile}; fi; elif [ -n -a ${target} = install ]; then echo === subversion-1.8.3 depends on package: ${subpkgfile} - not found; echo === USE_PACKAGE_DEPENDS_ONLY set - will not build from source; exit 1; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; else (cd $dir; make -DINSTALLS_DEPENDS $target $depends_args) ; fi; echo === Returning to build of subversion-1.8.3; fi ; fi ; done + set -e + lib=libdb-4.2.so + dir=/usr/ports/databases/db42 + target=install + depends_args='' + echo -n '=== subversion-1.8.3 depends on shared library: libdb-4.2.so' === subversion-1.8.3 depends on shared library: libdb-4.2.so+ found=0 + /bin/cat /usr/local/libdata/ldconfig/compat7x /usr/local/libdata/ldconfig/mysql /usr/local/libdata/ldconfig/portupgrade /usr/local/libdata/ldconfig/pth + dirs='/lib /usr/lib /usr/local/lib /usr/local/lib/compat /usr/local/lib/mysql /usr/local/lib/compat/pkg /usr/local/lib/pth' + test -f /lib/libdb-4.2.so + continue + test -f /usr/lib/libdb-4.2.so + continue + test -f /usr/local/lib/libdb-4.2.so + [ -x /usr/bin/file ] + file -b -L --mime-type /usr/local/lib/libdb-4.2.so + [ inode/symlink = application/x-sharedlib ] + continue + test -f /usr/local/lib/compat/libdb-4.2.so + continue + test -f /usr/local/lib/mysql/libdb-4.2.so + continue + test -f /usr/local/lib/compat/pkg/libdb-4.2.so + continue + test -f /usr/local/lib/pth/libdb-4.2.so + continue + [ 0 -eq 0 ] + echo ' - not found' - not found + echo '===Verifying for libdb-4.2.so in /usr/ports/databases/db42' ===Verifying for libdb-4.2.so in /usr/ports/databases/db42 So, file(1) call is the culprit here. Trying to reproduce in the console: [sobomax@pioneer ~]$ file -b -L --mime-type /usr/local/lib/libdb-4.2.so inode/symlink However: [sobomax@pioneer ~]$ hexdump -C /usr/local/lib/libdb-4.2.so | head -n 20 7f 45 4c 46 02 01 01 09 00 00 00 00 00 00 00 00 |.ELF| 0010 03 00 3e 00 01 00 00 00 d0 40 02 00 00 00 00 00 |...п@..| 0020 40 00 00 00 00 00 00 00 b8 a6 0e 00 00 00 00 00 |@...╦і..| 0030 00 00 00 00 40 00 38 00 05 00 40 00 1d 00 1a 00 |@.8...@ .| 0040 01 00 00 00 05 00 00 00 00 00 00 00 00 00 00 00 || 0050 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 || 0060 dc 45 0e 00 00 00 00 00 dc 45 0e 00 00 00 00 00 |эE..эE..| 0070 00 00 20 00 00 00 00 00 01 00 00 00 06 00 00 00 |.. .| 0080 00 50 0e 00 00 00 00 00 00 50 2e 00 00 00 00 00 |.P...P..| 0090 00 50 2e 00 00 00 00 00 fc 36 00 00 00 00 00 00 |.P..Э6..| 00a0 d8 3a 00 00 00 00 00 00 00 00 20 00 00 00 00 00 |ь: .| 00b0 02 00 00 00 06 00 00 00 e8 54 0e 00 00 00 00 00 |ХT..| 00c0 e8 54 2e 00 00 00 00 00 e8 54 2e 00 00 00 00 00 |ХT..ХT..| 00d0 a0 01 00 00 00 00 00 00 a0 01 00 00 00 00 00 00 |═...═...| 00e0 08 00 00 00 00 00 00 00 50 e5 74 64 04 00 00 00 |PЕtd| 00f0 5c 39 0d 00 00 00 00 00 5c 39 0d 00 00 00 00 00 |\9..\9..| 0100 5c 39 0d 00 00 00 00 00 b4 34 00 00 00 00 00 00 |\9..Є4..| 0110 b4 34 00 00 00 00
Re: ports/181913: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous
On Sep 8, 2013, at 08:14, O. Hartmann ohart...@zedat.fu-berlin.de wrote: On Sat, 7 Sep 2013 22:49:54 GMT rak...@freebsd.org wrote: Synopsis: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous State-Changed-From-To: open-patched State-Changed-By: rakuco State-Changed-When: Sat Sep 7 22:47:43 UTC 2013 State-Changed-Why: I don't think the previous version worked. From your description, it looks like you've switched to building with libc++ whereas libstdc++ was being used before. The upcoming Qt 4.8.5 plus a few patches which only made it to 4.8.6 (but we've backported) will finally make Qt build with libc++. We've just sent an exp-run request for Qt 4.8.5, and will hopefully fix all these errors once it is committed. http://www.freebsd.org/cgi/query-pr.cgi?pr=181913 I build the world/kernel since early this year with CXXFLAGS+= -stdlib=libc++ CXXFLAGS+= -std=c++11 in /etc/src.conf. I do not use those flags in /etc/make.conf! /etc/src.conf is supposed to target ONLY the /usr/src world, not the ports - this is as I interpret the man page for /etc/src.conf and it would be logical. But this rule/thinking seems to be broken by some includes from /usr/ports/Mk ingredients. Since r255321, -stdlib=libc++ is effectively the default, at least when you haven't set gcc as the default compiler. So it also applies to ports, which unavoidably will lead to a bit of fallout. My personal experience is that most C++-based ports compile fine with libc++ instead of libstdc++, except for a few that rely on internal libstdc++ details. However, -std=c++11 is *not* yet the default, and C++11 has different rules here and there, so some ports might fail to compile due to this. For some ports, too much hacking may be required to make them work with C++11. So in case of trouble, try removing -std=, or setting it to different values (c++0x, c++98, gnu++98, etc), to get the port to compile. Note the base system should have no problems with -std=c++11, so please continue to use the option in src.conf, and report any problems if you encounter them, so we can fix them. :-) -Dimitry signature.asc Description: Message signed with OpenPGP using GPGMail
Re: ports/181913: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous
On Sun, 8 Sep 2013 14:57:01 +0200 Dimitry Andric d...@freebsd.org wrote: On Sep 8, 2013, at 08:14, O. Hartmann ohart...@zedat.fu-berlin.de wrote: On Sat, 7 Sep 2013 22:49:54 GMT rak...@freebsd.org wrote: Synopsis: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous State-Changed-From-To: open-patched State-Changed-By: rakuco State-Changed-When: Sat Sep 7 22:47:43 UTC 2013 State-Changed-Why: I don't think the previous version worked. From your description, it looks like you've switched to building with libc++ whereas libstdc++ was being used before. The upcoming Qt 4.8.5 plus a few patches which only made it to 4.8.6 (but we've backported) will finally make Qt build with libc++. We've just sent an exp-run request for Qt 4.8.5, and will hopefully fix all these errors once it is committed. http://www.freebsd.org/cgi/query-pr.cgi?pr=181913 I build the world/kernel since early this year with CXXFLAGS+= -stdlib=libc++ CXXFLAGS+= -std=c++11 in /etc/src.conf. I do not use those flags in /etc/make.conf! /etc/src.conf is supposed to target ONLY the /usr/src world, not the ports - this is as I interpret the man page for /etc/src.conf and it would be logical. But this rule/thinking seems to be broken by some includes from /usr/ports/Mk ingredients. Since r255321, -stdlib=libc++ is effectively the default, at least when you haven't set gcc as the default compiler. So it also applies to ports, which unavoidably will lead to a bit of fallout. My personal experience is that most C++-based ports compile fine with libc++ instead of libstdc++, except for a few that rely on internal libstdc++ details. However, -std=c++11 is *not* yet the default, and C++11 has different rules here and there, so some ports might fail to compile due to this. For some ports, too much hacking may be required to make them work with C++11. So in case of trouble, try removing -std=, or setting it to different values (c++0x, c++98, gnu++98, etc), to get the port to compile. Note the base system should have no problems with -std=c++11, so please continue to use the option in src.conf, and report any problems if you encounter them, so we can fix them. :-) -Dimitry Hello Dimitry. I ONLY use -std=c++11 in /etc/src.conf. The base system had never problems so far since I use it. In /etc/make.conf, I avoid it. But, and this is obviously a logical incosistency, the ports system includes also /etc/src.conf, and I consider /etc/src.conf as base system only as the man page suggests. But the discussion has already been on the list. Somewhere in the basd.*.mk files, /etc/src.conf is included. And I guess therefore it comes to problems. Oliver signature.asc Description: PGP signature
[QAT] r326727: 4x leftovers
devel/py-mongoengine: update to 0.8.4 - Update to 0.8.4 - Use zip_safe=False - Remove leading article from COMMENT - Take maintainership Changes: http://docs.mongoengine.org/en/latest/changelog.html PR: ports/181486 Submitted by: wg (myself) Approved by:maintainer (timeout) - Build ID: 20130908135800-48145 Job owner: w...@freebsd.org Buildtime: 3 hours Enddate: Sun, 08 Sep 2013 17:12:53 GMT Revision: r326727 Repository: https://svnweb.freebsd.org/ports?view=revisionrevision=326727 - Port:devel/py-mongoengine 0.8.4 Buildgroup: 9.1-QAT/amd64 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908135800-48145-184176/py27-mongoengine-0.8.4.log Buildgroup: 9.1-QAT/i386 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908135800-48145-184177/py27-mongoengine-0.8.4.log Buildgroup: 8.4-QAT/amd64 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908135800-48145-184178/py27-mongoengine-0.8.4.log Buildgroup: 8.4-QAT/i386 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908135800-48145-184179/py27-mongoengine-0.8.4.log -- Buildarchive URL: https://qat.redports.org/buildarchive/20130908135800-48145 redports https://qat.redports.org/ ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Re: libiconv.a(iconv.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object
On Sun, Sep 08, 2013 at 12:52:03PM +0200, Jeremie Le Hen wrote: Hi, (Please Cc: me when replying, I'm not subscribed.) I created a fresh FreeBSD 9.1-RELEASE amd64 jail. No installed ports. I cannot build devel/gettext. Does anyone experience the same problem? I invariably run into the following problem: libtool: link: cc -shared -fPIC -DPIC .libs/bindtextdom.o .libs/dcgettext.o .libs/dgettext.o .libs/gettext.o .libs/finddomain.o .libs/hash-string.o .libs/loadmsgcat.o .libs/localealias.o .libs/textdomain.o .libs/l10nflist.o .libs/explodename.o .libs/dcigettext.o .libs/dcngettext.o .libs/dngettext.o .libs/ngettext.o .libs/plural.o .libs/plural-exp.o .libs/localcharset.o .libs/threadlib.o .libs/lock.o .libs/relocatable.o .libs/langprefs.o .libs/localename.o .libs/log.o .libs/printf.o .libs/setlocale.o .libs/version.o .libs/xsize.o .libs/osdep.o .libs/intl-compat.o -L/usr/local/lib /usr/local/lib/libiconv.a -O2 -Wl,-soname -Wl,libintl.so.9 -o .libs/libintl.so.9 /usr/bin/ld: /usr/local/lib/libiconv.a(iconv.o): relocation R_X86_64_32S against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/libiconv.a: could not read symbols: Bad value Full typescript here: http://people.freebsd.org/~jlh/typescript.gettext.txt For the record, Gabor Pali pointed that inside the jail, uname(1) still returned the version from my host (10-CURRENT), so the port makes wrong assumptions. This is solved using environement variables as described in the uname(1) manpage. -- Jeremie Le Hen Scientists say the world is made up of Protons, Neutrons and Electrons. They forgot to mention Morons. ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
[QAT] r326740: 4x leftovers
devel/py-mongoengine: take maintainership - Take maintainership (forgot to do so in the last commit) PR: ports/181486 Approved by:maintainer (timeout) - Build ID: 20130908150401-39362 Job owner: w...@freebsd.org Buildtime: 5 hours Enddate: Sun, 08 Sep 2013 20:32:59 GMT Revision: r326740 Repository: https://svnweb.freebsd.org/ports?view=revisionrevision=326740 - Port:devel/py-mongoengine 0.8.4 Buildgroup: 9.1-QAT/amd64 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908150401-39362-184400/py27-mongoengine-0.8.4.log Buildgroup: 9.1-QAT/i386 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908150401-39362-184401/py27-mongoengine-0.8.4.log Buildgroup: 8.4-QAT/amd64 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908150401-39362-184402/py27-mongoengine-0.8.4.log Buildgroup: 8.4-QAT/i386 Buildstatus: LEFTOVERS Log: https://qat.redports.org//~w...@freebsd.org/20130908150401-39362-184403/py27-mongoengine-0.8.4.log -- Buildarchive URL: https://qat.redports.org/buildarchive/20130908150401-39362 redports https://qat.redports.org/ ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org
Re: AFFECTS: 10-CURRENT users with any port depending on converters/libiconv
On Fri, 06 Sep 2013 09:34:52 +0200 Guido Falsi madpi...@freebsd.org wrote: On 09/06/13 05:16, AN wrote: Hi: I am posting to both lists because this problem affects users of current and ports, and I didn't know which would be more appropriate so please forgive me. # uname -a FreeBSD FBSD10 10.0-CURRENT FreeBSD 10.0-CURRENT #80 r255129: Sun Sep 1 16:01:36 CDT 2013 root@FBSD10:/usr/obj/usr/src/sys/MYKERNEL amd64 I am trying to update my ports following the entry in updating, but it does not seem to be working correctly. I followed the directions exactly, and after 30 mins this is what has happened: # cat ports_to_update | xargs portupgrade -vf --- Session started at: Thu, 05 Sep 2013 21:12:10 -0500 [Reading data from pkg(8) ... - 890 packages found - done] Shared object libiconv.so.3 not found, required by httpd make: /usr/ports/Mk/bsd.apache.mk line 278: warning: Couldn't read shell's output for /usr/local/sbin/httpd -V | /usr/bin/sed -ne 's/^Server version: Apache\/\([0-9]\)\.\([0-9]*\).*/\1\2/p' Shared object libiconv.so.3 not found, required by httpd This is bsd.apache.mk trying to get the apache version. but the apache's httpd binary cannot run because it can't find libiconv.so.3. apxs:Error: Sorry, no shared object support for Apache. apxs:Error: available under your platform. Make sure. apxs:Error: the Apache module mod_so is compiled into. apxs:Error: your server binary `/usr/local/sbin/httpd'.. make: /usr/ports/Mk/bsd.apache.mk line 284: warning: /usr/local/sbin/apxs -q MPM_NAME returned non-zero status ** Port marked as IGNORE: www/mod_dnssd: is marked as broken: : Error from bsd.apache.mk. apache is installed (or APACHE_PORT is defined) and port requires apache22 at least Here is what I have done: # pkg query %ro libiconv ports_to_update [root@FBSD10 ~]# cat ports_to_update ...lots of output # pkg delete -f libiconv pkg: You are trying to delete package(s) which has dependencies that are still required: ... delete these packages anyway in forced mode Deinstallation has been requested for the following 1 packages: libiconv-1.14_1 The deinstallation will free 2 MB Proceed with deinstalling packages [y/N]: y [1/1] Deleting libiconv-1.14_1... deleting anyway done Now the update process is stuck here: ** Port marked as IGNORE: www/mod_dnssd: is marked as broken: : Error from bsd.apache.mk. apache is installed (or APACHE_PORT is defined) and port requires apache22 at least there are 2 ruby processes running for a long time, but nothing is happening to the update. 43998 root520 64912K 33368K piperd 5 2:21 5.96% ruby19{ruby19} 43998 root520 64912K 33368K select 1 0:00 5.96% ruby19{ruby19} So, it seems my system is broken now. Did I do something wrong? How can the upgrade work if so many ports depend on iconv? What should I do now? Should I reinstall libiconv? Good news is the update process did not really update anything, judging from the output you sent. If you just reinstall libiconv everything should go back to how it was, at least you get a working system. I admit I did not foresee this condition arising when I wrote the instructions, here is a modified procedure you can follow and report back about, so I can modify the UPDATING entry: # pkg query %ro libiconv ports_to_update # cp /usr/local/lib/libiconv.so.3 /usr/local/lib/compat/pkg/ # ldconfig -R (1) # pkg delete -f libiconv # cat ports_to_update | xargs portupgrade -f (1) not sure if ldconfig -R is really needed, but It will not do any harm I added the step to preserve libiconv.so.3 in /usr/local/lib/compat/pkg which is in the default library search path. In this way libiconv and it's include file shouldn't be found by configure scripts and the like and they should link to the system one, while existing binaries should keep working linking to the preserved one in lib/compat. Any help is appreciated. I hope this helps you, just ask for any clarifications and further help as needed on this matter. Just for the record: after three days of cleaning up this mess today two boxes got ready, servers, without any GUI or fancy X11 stuff. They work. Two other development boxes reject compiling kdelibs and stuff. After the final update today on CURRENT r255398, ports like kdevelop, firefox, libreoffice(!) and others crash along with virtualbox-ose-kmod (coredump). Most of that stuff, including libreoffice, firefox was ready last night, including kdelibs and the stuff for kdevelop. It worked today - until now, r255398. I do not know what happened here, but I'm through with it. Two obviously independend processes, one from the ports side, another from the OS side, intertwined and messy as hell rendered the systems completely unusable. As I reported on several occasions, after r255259 it wasn't
Re: ports/181913: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous
On Sun, 8 Sep 2013 14:57:01 +0200 Dimitry Andric d...@freebsd.org wrote: On Sep 8, 2013, at 08:14, O. Hartmann ohart...@zedat.fu-berlin.de wrote: On Sat, 7 Sep 2013 22:49:54 GMT rak...@freebsd.org wrote: Synopsis: devel/qt4-script: /usr/include/c++/v1/type_traits:3175:22: error: call to 'swap' is ambiguous State-Changed-From-To: open-patched State-Changed-By: rakuco State-Changed-When: Sat Sep 7 22:47:43 UTC 2013 State-Changed-Why: I don't think the previous version worked. From your description, it looks like you've switched to building with libc++ whereas libstdc++ was being used before. The upcoming Qt 4.8.5 plus a few patches which only made it to 4.8.6 (but we've backported) will finally make Qt build with libc++. We've just sent an exp-run request for Qt 4.8.5, and will hopefully fix all these errors once it is committed. http://www.freebsd.org/cgi/query-pr.cgi?pr=181913 I build the world/kernel since early this year with CXXFLAGS+= -stdlib=libc++ CXXFLAGS+= -std=c++11 in /etc/src.conf. I do not use those flags in /etc/make.conf! /etc/src.conf is supposed to target ONLY the /usr/src world, not the ports - this is as I interpret the man page for /etc/src.conf and it would be logical. But this rule/thinking seems to be broken by some includes from /usr/ports/Mk ingredients. Since r255321, -stdlib=libc++ is effectively the default, at least when you haven't set gcc as the default compiler. So it also applies to ports, which unavoidably will lead to a bit of fallout. My personal experience is that most C++-based ports compile fine with libc++ instead of libstdc++, except for a few that rely on internal libstdc++ details. However, -std=c++11 is *not* yet the default, and C++11 has different rules here and there, so some ports might fail to compile due to this. For some ports, too much hacking may be required to make them work with C++11. So in case of trouble, try removing -std=, or setting it to different values (c++0x, c++98, gnu++98, etc), to get the port to compile. Note the base system should have no problems with -std=c++11, so please continue to use the option in src.conf, and report any problems if you encounter them, so we can fix them. :-) -Dimitry Hello Dimitry, btw, see PR ports/181932. This is definitely NOT libc++ related. It came up since nearly all qt4-related clients (also kdelibs) fail and drop core on r255398 - they worked prior to the last update today. I tried recompiling qt4- and kdelibs4 to get my kdevelop environment as well as libreoffice back (the drop core, as well as firefox, out of the blue). I also tried compiling those ports without any settings of CXXFLAGS in /etc/src.conf, but it doens't help. I can not understand why two critical changes from different branches of the maintainig get the same time into the public (iconv/ports and libstdc++ vanishing). Maybe I'm wrong here, but after three days, two nights non-stop updating I'm through with this toy. signature.asc Description: PGP signature
Re: x11/kdelibs4: /usr/local/include/grantlee/typeaccessor.h:70:40: error: no matching function for call to 'distance' return
Raphael Kubo da Costa rak...@freebsd.org writes: O. Hartmann ohart...@zedat.fu-berlin.de writes: After a messy iconv orgy update session, I update yesterday x11/kdelib4 successfully. After updating the OS to FreeBSD 10.0-CURRENT #2 r255356: Sat Sep 7 13:04:03 CEST 2013 amd64 today, I run today surprisingly into this error: [...] In file included from /usr/ports/x11/kdelibs4/work/kdelibs-4.10.5/kdeui/tests/proxymodeltestsuite/modeleventlogger.cpp:33: In file included from /usr/local/include/grantlee_core.h:24: In file included from /usr/local/include/grantlee_templates.h:34: In file included from /usr/local/include/grantlee/metatype.h:27: /usr/local/include/grantlee/typeaccessor.h:70:40: error: no matching function for call to 'distance' return QVariant::fromValueint( std::distance( container.begin(), container.end() ) ); [...] grantlee has also been update successfully, I recompiled it today again successfulyy, but without any success. Are you using libc++ and is the error message much bigger than that? I've fixed that one in Qt and the patch is part of 4.8.5. I'll see how much work is left to bring in 4.8.5 into ports. Fixed in r326778. ___ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to freebsd-ports-unsubscr...@freebsd.org