Re: rpm hangs while building some packages under solaris
the first two commands also hang. the last command hangs after the sed (last) command. I have attached my console output. On Fri, 2004-03-19 at 03:21, Ralf S. Engelschall wrote: On Thu, Mar 18, 2004, Doug Henry wrote: Has anyone experienced a problem under solaris (I'm using 8) where rpm will hang when doing a rpm --rebuild package? If I remove the lines that use {platform_t} from the spec file I can build it. The first package in the build sequence that I experience this is binutils. It happens using openpkg versions 1.3 and 2.0. Any info on debugging/fixing this problem would be appreciated. If %{l_platform -t} hangs, what does also hang the following commands? $ prefix/bin/openpkg rpm --eval '%{l_platform -t}' $ prefix/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st And if yes, try to execute $ sh -x prefix/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st and tell us the output, please. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED] Script started on Fri 19 Mar 2004 07:53:50 AM EST $ openpkg rpm --eval '%{l_platform -t}' ^C$ $ /home/losradar/tools/SunOS-5.8/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st ^C $ $ sh -x /home/losradar/tools/SunOS-5.8/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st + [ 9 -eq 0 ] + [ .platform = .-h ] + [ .platform = .--help ] + [ .platform = .-v ] + [ .platform = .--version ] + [ .platform = .-r ] + [ .platform = .--recreate ] + [ .platform = .-d ] + [ .platform = .--debug ] + echo /home/losradar/tools/SunOS-5.8/lib/openpkg/shtool + sed -e s;.*/\([^/]*\)$;\1; -e s;-sh$;; -e s;\.sh$;; name=shtool tool=platform + shift arg_spec= opt_spec= gen_tmpfile=no str_tool=platform str_usage=[-F|--format format] [-S|--sep string] [-C|--conc string] [-L|--lower] [-U|--upper] [-v|--verbose] [-c|--concise] [-n|--no-newline] [-t|--type type] [-V|--version] [-h|--help] arg_spec=0= opt_spec=F:S:C:L.U.v.c.n.t:d.V.h. opt_alias=F:format,S:sep,C:conc,L:lower,U:upper,v:verbose,c:consise,t:type,n:no-newline,V:version,h:help opt_F=%{sp} (%{ap}) opt_S= opt_C=/ opt_L=no opt_U=no opt_t= opt_v=no opt_c=no opt_n=no opt_V=no opt_h=no ASC_TAB= ASC_NL= + [ .platform != . ] toolcmd=/home/losradar/tools/SunOS-5.8/lib/openpkg/shtool platform toolcmdhelp=shtool platform msgprefix=shtool:platform + echo 0= + sed -e s/^\([0-9]*\)\([+=]\)/arg_NUMS=\1; arg_MODE=\2/ + eval arg_NUMS=0; arg_MODE== arg_NUMS=0 arg_MODE== + echo h.F:S:C:L.U.v.c.n.t:d.V.h. + sed -e s/\([a-zA-Z0-9]\)\([.:+]\)/opt_MODE_\1=\2;/g + eval opt_MODE_h=.;opt_MODE_F=:;opt_MODE_S=:;opt_MODE_C=:;opt_MODE_L=.;opt_MODE_U=.;opt_MODE_v=.;opt_MODE_c=.;opt_MODE_n=.;opt_MODE_t=:;opt_MODE_d=.;opt_MODE_V=.;opt_MODE_h=.; opt_MODE_h=. opt_MODE_F=: opt_MODE_S=: opt_MODE_C=: opt_MODE_L=. opt_MODE_U=. opt_MODE_v=. opt_MODE_c=. opt_MODE_n=. opt_MODE_t=: opt_MODE_d=. opt_MODE_V=. opt_MODE_h=. + echo h:help,F:format,S:sep,C:conc,L:lower,U:upper,v:verbose,c:consise,t:type,n:no-newline,V:version,h:help + tr x- x_ + sed -e s/\([a-zA-Z0-9]\):\([^,]*\),*/opt_ALIAS_\2=\1;/g ^C $ $ exit script done on Fri 19 Mar 2004 07:55:05 AM EST
Re: rpm hangs while building some packages under solaris
Forgot to mention in my other message that it seems to be a more general problem than just l_platform. The same thing happens when I try to build ncurses and I don't think there are any platform commands in that spec file. Thanks On Fri, 2004-03-19 at 03:21, Ralf S. Engelschall wrote: On Thu, Mar 18, 2004, Doug Henry wrote: Has anyone experienced a problem under solaris (I'm using 8) where rpm will hang when doing a rpm --rebuild package? If I remove the lines that use {platform_t} from the spec file I can build it. The first package in the build sequence that I experience this is binutils. It happens using openpkg versions 1.3 and 2.0. Any info on debugging/fixing this problem would be appreciated. If %{l_platform -t} hangs, what does also hang the following commands? $ prefix/bin/openpkg rpm --eval '%{l_platform -t}' $ prefix/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st And if yes, try to execute $ sh -x prefix/lib/openpkg/shtool platform -n -L -S -C + -F %at-%st and tell us the output, please. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED] __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED]
SGI support
I saw that IRIX is in the forecasted category on the openpkg website. Is there a group working on that port yet? I have access to several irix machines and am willing to test and build if it will help that effort. Preliminary testing suggests there are a few rpm/cpio problems to work out. Let me know if I can help out. Thanks, Doug Henry [EMAIL PROTECTED] Software Development Northrop Grumman __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED]
thread safe sqlite
sqlite can be built to be thread safe. In keeping with the current with_threads standard, I have attached a patch that adds this functionality to the sqlite.spec file. *** sqlite.spec Mon Mar 8 13:44:08 2004 --- /tmp/sqlite.spec.new Thu Apr 22 08:54:15 2004 *** Release: 20040308 *** 40,45 --- 40,46 %option with_utf8no %option with_assert no %option with_readlineno + %option with_threads no # list of sources Source0: http://www.hwaci.com/sw/sqlite/sqlite-%{version}.tar.gz *** AutoReqProv: no *** 98,103 --- 99,107 config_TARGET_READLINE_LIBS=%{l_ldflags} -lreadline -ltermcap export config_TARGET_READLINE_LIBS %endif + %if %{with_threads} == yes + CFLAGS=$CFLAGS -DTHREADSAFE=1 + %endif ./configure \ --prefix=%{l_prefix} \ %if %{with_utf8} == yes
Re: Format of ~/.openpkg/build
Maybe format is not a problem, what does that stuff mean. I can't find explanation anywhere. My build file is: [/server/tools/bin/openpkg rpm] -Dgcc::with_f77=yes -Dgcc::with_bounds=yes -Dglib::with_threads=yes On Wed, 2004-04-28 at 12:41, Ralf S. Engelschall wrote: On Wed, Apr 28, 2004, Doug Henry wrote: did anyone ever figure out what the correct format is? documentation? this seems like a VERY good thing to know if you are building more than one openpkg system from source. The syntax is what was mentioned. For instance, my ~/.openpkg/build currently reads: - [/usr/opkg/bin/openpkg rpm] -X -r /e/openpkg/SRC -f /e/openpkg/SRC/00INDEX.rdf.bz2 -P sudo -N sudo -E j2se -E uvscan -E kde-qt -E mozilla - Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED] __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED]
rpm/db open problems
I am trying to install openpkg on an NFS mount and I'm having problems. Do the following errors look like something that has been worked on by openpkg before? rpmdb: mmap: resource temporarily unavailable error: db4 error(11) from dbenv-open: resource temporarily unavailable cannot open packages index using db3 cannot open packages database in an initial google query tells me that this is a fairly common rpm/db issue, but I haven't seem any real solution. Thanks __ The OpenPKG Projectwww.openpkg.org User Communication List [EMAIL PROTECTED]
mathomatic problem
I think there is a small problem with the mathomatic spec file. Mathomatic requires ncurses, but has -lcurses in the LIBS line of spec file. I believe this needs to be -lncurses. -doug
Re: gcc40 f77 problem
I didn't see that package, I noticed the same problem with gcc41, so I'll try to build it with mpfr instead of gmp. On 7/21/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Thu, Jul 21, 2005, Doug Henry wrote: It seems the f77 support has changed to f95 support in gcc40.Setting with_f77 will fail when building this package. Here are the changes I made to make it work: - Changed f77 to f95 in the enable-languages option - When f95 option is used the gmp package becomes a build requirement - The default gmp package does not have MPFR support, so I modified gmp to have that option (--enable-mpfr) - gcc40 cannot find gmp.h, so I added %{l_cppflags} the the CFLAGS listQuestion: we have a mpfr package in CURRENT. Does this not work withGCC? My understand is that the --enable-mpfr option to GMP just enables a compatibility API for MPFR. Hence I ask myself whether we bettershould use the original one instead... Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: gcc40 f77 problem
gmp/mpfr seems to work fine. On 7/21/05, Doug Henry [EMAIL PROTECTED] wrote: I didn't see that package, I noticed the same problem with gcc41, so I'll try to build it with mpfr instead of gmp. On 7/21/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Thu, Jul 21, 2005, Doug Henry wrote: It seems the f77 support has changed to f95 support in gcc40.Setting with_f77 will fail when building this package. Here are the changes I made to make it work: - Changed f77 to f95 in the enable-languages option - When f95 option is used the gmp package becomes a build requirement - The default gmp package does not have MPFR support, so I modified gmp to have that option (--enable-mpfr) - gcc40 cannot find gmp.h, so I added %{l_cppflags} the the CFLAGS listQuestion: we have a mpfr package in CURRENT. Does this not work withGCC? My understand is that the --enable-mpfr option to GMP just enables a compatibility API for MPFR. Hence I ask myself whether we bettershould use the original one instead... Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: http://wiki.openpkg.org/?BuildingOpenPKGunderAIX
I have a subset of openpkg-current running on two of my IRIX octane 2 systems. If something gets going count me in for development/testing/building. I have octane and O2's available. On 7/27/05, Matthias Kurz [EMAIL PROTECTED] wrote: On Wed, Jul 27, 2005, Doug Summers wrote: Matthias Kurz wrote:{...] But then. Did you try openpkg build ? Did you experience any problems ?[...] Nope, never tried it. Try it. I think this is the direction, we should go. Even when we needsome improvements. That's my next test on a fresh Solaris 8 system. My head's kind of swimming right now with all of the platforms I'm trying to get OpenPKG to install on - so far only a no-go on AIX 5.2, Irix, Tru64, and HPUX 11.23/IA64.Hey, you're using IRIX ? Maybe i'm going to collaborate, there. It'snot the most used system, here. But at least i do have some resources. (mk)--Matthias Kurz; Fuldastr. 3; D-28199 Bremen; VOICE +49 421 53 600 47 Im prämotorischen Cortex kann jeder ein Held sein. (bdw) __The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: OpenPKG 2.4.2 and SGI IRIX 6.5
I have openpkg-2.2.1 (I'll have to verify that version) running on my IRIX boxes. I then installed many of the packages from 2.4/CURRENT on top of that. All of the newer openpkg releases have rpm build problems. I haven't looked in to much more than that because its been working pretty well. When I get back to work on Monday I can upload my rpm -qa if you want to see which packages I have been able to get working. I do remember perl causing some problems, I ended up building it from source. On 8/4/05, Doug Summers [EMAIL PROTECTED] wrote: Getting really close to getting this to build, using SGI's pre-built gcc3, which returns:Configured with: ../configure --prefix=/usr/freeware--enable-version-specific-runtime-libs --disable-shared --enable-threads --enable-haifa --enable-libgcj --disable-c-mbcharThread model: singlegcc version 3.3I've found some references to this problem from people trying to getOpenSSH compiled:/tmp/build/usr/freeware/bin/gcc -DHAVE_CONFIG_H -I. -I. -I.. -I. -I.. -I../popt -DOPENPKG -DOPENPKG_IRIX64 -I/tmp/openpkg-2.4.2/zlib-1.2.3-I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2-DOPENPKG -DOPENPKG_IRIX64 -I/tmp/openpkg-2.4.2/zlib-1.2.3-I/tmp/openpkg- 2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2 -O2-D_GNU_SOURCE -D_REENTRANT -MT stubs.lo -MD -MP -MF .deps/stubs.Tpo -cstubs.c -o stubs.oIn file included from stubs.c:10:../misc/basename.c:10: error: conflicting types for `basename' /usr/include/libgen.h:36: error: previous declaration of `basename'make[2]: *** [stubs.lo] Error 1make[2]: Leaving directory `/tmp/openpkg-2.4.2/rpm-4.2.1/rpmio'make[1]: *** [all-recursive] Error 1make[1]: Leaving directory `/tmp/openpkg- 2.4.2/rpm-4.2.1'make: *** [all] Error 2__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
feature request
I was wondering if we could put in an option for gcc (gcc41 would be ok) to enable AWT when java is built. This could be coupled with the existing java flag or a new option (with_gtkawt). The preferred flag would be: --enable-java-awt=gtk,xlib This would of course add gtk as a dependency for gcc::with_java which is a good reason to make it an extra option. Thanks! -doug
Re: OpenPKG 2.4.2 and SGI IRIX 6.5
here is the gcc spec output and the packages I have built on my octane. sh-2.05b$ gcc -v Reading specs from /local/tools/bin/../lib/gcc/mips-sgi-irix6.5/3.4.4/specs Configured with: ../configure --cache-file=./config.cache --srcdir=/local/tools/IRIX64-6.5/RPM/TMP/gcc-3.4.4/obj/.. --prefix=/local/tools/IRIX64-6.5 --exec-prefix=/local/tools/IRIX64-6.5 --includedir=/local/tools/IRIX64-6.5/include/gcc --libexecdir=/local/tools/IRIX64-6.5/libexec/gcc --with-gxx-include-dir=/local/tools/IRIX64-6.5/include/g++ --with-local-prefix=/local/tools/IRIX64-6.5/lib/gcc --enable-languages=c,c++,f77 --enable-threads=posix --disable-maintainer-mode --disable-shared --disable-nls --with-gnu-ld --with-ld=/local/tools/IRIX64-6.5/bin/ld --with-gnu-as --with-as=/local/tools/IRIX64-6.5/bin/as Thread model: posix gcc version 3.4.4 (OpenPKG-2.4) sh-2.05b$ openpkg rpm -qa | sort astyle-cvs-20050618 autoconf-2.59-20040429 automake-1.9.6-20050711 bash-3.0.16-20050727 binutils-2.16.1-20050613 bison-2.0-20050209 bzip2-1.0.3-20050506 clo++-0.6.4-20050406 coreutils-5.2.1-20050111 cpptest-1.0.2-20050325 cvs-1.12.11-20050612 doxygen-1.4.3-2.4.0 epm-3.7-20041109 expat-1.95.8-20050113 flex-2.5.31-20050216 fontconfig-2.2.90-20050212 fox-1.4.16-20050613 freetype-2.1.9-20050707 fsl-1.6.0-2.4.0 gawk-3.1.4-20040824 gcc-3.4.4-2.4.0 gcc33-3.3.6-20050503 gd-2.0.33-20050615 gdb-6.3-20041210 gettext-0.14.1-2.3.0 glib-1.2.10-20040428 grep-2.5.1a-20050228 gtk-1.2.10-20040920 gzip-1.3.5-20050724 infozip-2.31.5.52-20050726 jpeg-6b-20041217 libiconv-1.9.2-20041022 libxml-2.6.20-20050711 libxslt-1.1.14-20050403 lout-3.30-20041028 m4-1.4.3-20050331 make-3.80-20040207 nano-1.3.8-20050701 ncurses-5.4.20050723-20050724 openpkg-2.2.1-2.2.1 openpkg-tools-0.8.37-2.4.0 orbit-0.5.17-20040207 patch-2.5.9-20050606 pcre-6.1-20050622 pkgconfig-0.18-20050628 png-1.2.8-20050216 ptypes-2.0.2-20050414 readline-5.0.5-20050617 sharutils-4.3.81-20050625 slang-1.4.9-20040207 sqlite2-2.8.16-20050325 sqlite3-3.2.1-20050512 tar-1.15.1-20050131 texinfo-4.8-20050104 tiff-3.7.3-20050706 vtk-4.4.2-20050325 x11-0-20040809 xmlwrapp-0.5.0-20050325 zlib-1.2.3-20050722 sh-2.05b$ exit On 8/5/05, Doug Henry [EMAIL PROTECTED] wrote: I have openpkg-2.2.1 (I'll have to verify that version) running on my IRIX boxes. I then installed many of the packages from 2.4/CURRENT on top of that. All of the newer openpkg releases have rpm build problems. I haven't looked in to much more than that because its been working pretty well. When I get back to work on Monday I can upload my rpm -qa if you want to see which packages I have been able to get working. I do remember perl causing some problems, I ended up building it from source. On 8/4/05, Doug Summers [EMAIL PROTECTED] wrote: Getting really close to getting this to build, using SGI's pre-built gcc3, which returns:Configured with: ../configure --prefix=/usr/freeware--enable-version-specific-runtime-libs --disable-shared --enable-threads --enable-haifa --enable-libgcj --disable-c-mbcharThread model: singlegcc version 3.3I've found some references to this problem from people trying to getOpenSSH compiled:/tmp/build/usr/freeware/bin/gcc -DHAVE_CONFIG_H -I. -I. -I.. -I. -I.. -I../popt -DOPENPKG -DOPENPKG_IRIX64 -I/tmp/openpkg-2.4.2/zlib-1.2.3-I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2-DOPENPKG -DOPENPKG_IRIX64 -I/tmp/openpkg-2.4.2/zlib-1.2.3-I/tmp/openpkg- 2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2 -O2-D_GNU_SOURCE -D_REENTRANT -MT stubs.lo -MD -MP -MF .deps/stubs.Tpo -cstubs.c -o stubs.oIn file included from stubs.c:10:../misc/basename.c:10: error: conflicting types for `basename' /usr/include/libgen.h:36: error: previous declaration of `basename'make[2]: *** [stubs.lo] Error 1make[2]: Leaving directory `/tmp/openpkg-2.4.2/rpm-4.2.1/rpmio'make[1]: *** [all-recursive] Error 1make[1]: Leaving directory `/tmp/openpkg- 2.4.2/rpm-4.2.1'make: *** [all] Error 2__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
bootstrap development
what is the procedure for extracting and re-assembling the bootstrap package? I would like to experiment with adding in a few patches and then making a new bootstrap script. -doug
Re: OpenPKG 2.4.2 and SGI IRIX 6.5
This error is very common under irix: /usr/freeware/bin/gcc -DHAVE_CONFIG_H -I. -I. -I.. -I. -I.. -I../popt -DOPENPKG -DOPENPKG_IRIX -I/tmp/openpkg-2.4.2/zlib-1.2.3 -I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2 -DOPENPKG -DOPENPKG_IRIX -I/tmp/openpkg-2.4.2/zlib-1.2.3 -I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2 -O2 -D_GNU_SOURCE -D_REENTRANT -MT fts.lo -MD -MP -MF .deps/fts.Tpo -c fts.c -o fts.o fts.c: In function `Fts_open': fts.c:181: error: invalid type argument of `unary *' fts.c:222: error: invalid type argument of `unary *' fts.c:231: error: invalid type argument of `unary *' fts.c: In function `Fts_close': fts.c:371: error: invalid type argument of `unary *' fts.c: In function `Fts_read': fts.c:545: error: invalid type argument of `unary *' fts.c:568: error: invalid type argument of `unary *' fts.c: In function `Fts_set': fts.c:595: error: invalid type argument of `unary *' fts.c: In function `Fts_children': fts.c:611: error: invalid type argument of `unary *' fts.c:624: error: invalid type argument of `unary *' fts.c: In function `fts_build': fts.c:825: error: invalid type argument of `unary *' fts.c:849: error: invalid type argument of `unary *' fts.c: In function `fts_stat': fts.c:987: error: invalid type argument of `unary *' fts.c: In function `fts_palloc': fts.c:1152: error: invalid type argument of `unary *' fts.c: In function `fts_safe_changedir': fts.c:1227: error: invalid type argument of `unary *' fts.c:1238: error: invalid type argument of `unary *' make[2]: *** [fts.lo] Error 1 make[2]: Leaving directory `/tmp/openpkg-2.4.2/rpm-4.2.1/rpmio' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/tmp/openpkg-2.4.2/rpm-4.2.1' make: *** [all] Error 2 + exit 2 runscript:ERROR: script returned non-null value On 8/9/05, Doug Henry [EMAIL PROTECTED] wrote: Nope the 2.2.1 version just happened to work. All others I have tried fail while building rpm. I think the basename issue is just masking a bigger problem. From googling other errors I have seen, rpm needs to have some system specific work done to it in many cases. I will try sometime to build a newer openpkg and post the other errors I have seen. If I remember correctly building openpkg from a working openpkg yields different results from a bootstrap build. Also, the errors on my O2 vs. Octane are quite different, even though the compilers are the same. On 8/9/05, Doug Summers [EMAIL PROTECTED] wrote: Doug Henry wrote: here is the gcc spec output and the packages I have built on my octane. sh-2.05b$ gcc -v Reading specs from /local/tools/bin/../lib/gcc/mips-sgi-irix6.5/3.4.4/specs Configured with: ../configure --cache-file=./config.cache --srcdir=/local/tools/IRIX64-6.5/RPM/TMP/gcc-3.4.4/obj/.. --prefix=/local/tools/IRIX64-6.5 --exec-prefix=/local/tools/IRIX64-6.5 --includedir=/local/tools/IRIX64- 6.5/include/gcc --libexecdir=/local/tools/IRIX64-6.5/libexec/gcc --with-gxx-include-dir=/local/tools/IRIX64-6.5/include/g++ --with-local-prefix=/local/tools/IRIX64-6.5/lib/gcc --enable-languages=c,c++,f77 --enable-threads=posix --disable-maintainer-mode --disable-shared --disable-nls --with-gnu-ld --with-ld=/local/tools/IRIX64-6.5/bin/ld --with-gnu-as --with-as=/local/tools/IRIX64-6.5/bin/as Thread model: posix gcc version 3.4.4 (OpenPKG-2.4) sh-2.05b$ openpkg rpm -qa | sort astyle-cvs-20050618 autoconf-2.59-20040429 automake-1.9.6-20050711 bash-3.0.16-20050727 binutils-2.16.1-20050613 bison-2.0-20050209 bzip2-1.0.3-20050506 clo++-0.6.4-20050406 coreutils-5.2.1-20050111 cpptest-1.0.2-20050325 cvs-1.12.11-20050612 doxygen-1.4.3-2.4.0 epm-3.7-20041109 expat-1.95.8-20050113 flex-2.5.31-20050216 fontconfig-2.2.90-20050212 fox-1.4.16-20050613 freetype-2.1.9-20050707 fsl-1.6.0-2.4.0 gawk-3.1.4-20040824 gcc-3.4.4-2.4.0 gcc33-3.3.6-20050503 gd-2.0.33-20050615 gdb-6.3-20041210 gettext-0.14.1-2.3.0 glib-1.2.10-20040428 grep-2.5.1a-20050228 gtk-1.2.10-20040920 gzip-1.3.5-20050724 infozip-2.31.5.52-20050726 jpeg-6b-20041217 libiconv-1.9.2-20041022 libxml-2.6.20-20050711 libxslt-1.1.14-20050403 lout-3.30-20041028 m4-1.4.3-20050331 make-3.80-20040207 nano-1.3.8-20050701 ncurses-5.4.20050723-20050724 openpkg-2.2.1-2.2.1 openpkg-tools-0.8.37-2.4.0 orbit-0.5.17-20040207 patch-2.5.9-20050606 pcre-6.1-20050622 pkgconfig-0.18-20050628 png-1.2.8-20050216 ptypes-2.0.2-20050414 readline-5.0.5-20050617 sharutils-4.3.81-20050625 slang-1.4.9-20040207 sqlite2-2.8.16-20050325 sqlite3-3.2.1-20050512 tar-1.15.1-20050131 texinfo-4.8-20050104 tiff-3.7.3-20050706 vtk-4.4.2-20050325 x11-0-20040809 xmlwrapp-0.5.0-20050325 zlib-1.2.3-20050722 sh-2.05b$ exit On 8/5/05, *Doug Henry* [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: I have openpkg-2.2.1 (I'll have to verify that version) running on my IRIX boxes.I then installed many of the packages from 2.4/CURRENT on top of that.All of the newer openpkg releases have rpm build problems.I haven't looked in to much more than that because its been working
wx depends problem
the wx package in CURRENT has glib/gtk listed as required. The default (as of 2.6 I think) is to use gtk2, so if gtk and gtk2 are installed, wx will be built using gtk2 but the requires will remain gtk.
openpkg-2.4 t1lib problem
it seems t1lib needs the xaw3d package but it not required. t1lib does not seem to use the CPPFLAGS variable correctly, so I had to add %{l_cppflags} to CFLAGS and things seem to compile. there is a link error because it is trying to link to Xaw and Xaw3d is the installed library name. after making a symlink from libXaw3d.a to libXaw.a all is well.
syslog config (a bit off topic)
Since this is the one list I can think of that probably has a lot of administrator participants I wanted to through this question out. If I want to log attempts to change the system time on a system (linux in particular) where would be a good place to find logging/auditing capabilities? Using audit software included with solaris and irix I can accomplish these tasks easily, but linux does not seem so nice. I am currently using syslog-ng under linux, and with all messages/levels being logged I can't seem to get a message about failed attempts to change system time. Most standard things like failed logins/su show up, but a permission denied when trying rm /usr/bin/passwd does not. Feel free to contact me offline if you wish to prevent traffic to this list. I guess I'm basically trying to figure out what auditing capabilities I have, and if I need more, where (if) I can get them. Thanks for any input. -doug
Re: OpenPKG 2.4.2 and SGI IRIX 6.5 - FIXED
I have openpkg-2.4.2 running on my octane2, and I'm almost embarrassed at how easy the fix is. I extracted the openpkg.sh file using the command: sh openpkg.sh -t | tar xvf - Then changed the following line in openpkg.spec: line 765: LIBS=$LDFLAGS -lz -lbz2 -lbeecrypt line 765: LIBS=$LDFLAGS -lz -lbz2 -lbeecrypt -lgen and then issued ./openpkg.boot --prefix and I have now been thanked for flying openpkg. I believe what happened is the configure script for rpm detected libgen.h has a basename function, so it added that to the system.h header to be included. But when the config script sample program was compiled it failed to link because -lgen was left off, so it tried to compile the misc/basename.c file included with rpm. This of course causes a multiple declaration error for basename and fails. Next on my list is to try on my O2 because I think I have more/different errors on that box for some reason. Then I will finish up by trying the openpkg in CURRENT. There are many packages that need tweaked to work on IRIX, but if the bootstrap works we should have a better change at ironing those problems out. FYI, one of the first big problems you will run in to is perl, more on that later. Later. -doug On 8/13/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Wed, Aug 10, 2005, Doug Summers wrote: Doug Henry wrote: This error is very common under irix: /usr/freeware/bin/gcc -DHAVE_CONFIG_H -I. -I. -I.. -I. -I.. -I../popt -DOPENPKG -DOPENPKG_IRIX -I/tmp/openpkg- 2.4.2/zlib-1.2.3 -I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg-2.4.2/beecrypt-4.1.2 -DOPENPKG -DOPENPKG_IRIX -I/tmp/openpkg-2.4.2/zlib-1.2.3 -I/tmp/openpkg-2.4.2/bzip2-1.0.3 -I/tmp/openpkg- 2.4.2/beecrypt-4.1.2 -O2 -D_GNU_SOURCE -D_REENTRANT -MT fts.lo -MD -MP -MF .deps/fts.Tpo -c fts.c -o fts.o fts.c: In function `Fts_open': fts.c:181: error: invalid type argument of `unary *' fts.c:222: error: invalid type argument of `unary *' fts.c:231: error: invalid type argument of `unary *' fts.c: In function `Fts_close': fts.c:371: error: invalid type argument of `unary *' fts.c: In function `Fts_read': fts.c:545: error: invalid type argument of `unary *' fts.c:568: error: invalid type argument of `unary *' fts.c: In function `Fts_set': fts.c:595: error: invalid type argument of `unary *' fts.c: In function `Fts_children': fts.c:611: error: invalid type argument of `unary *' fts.c:624: error: invalid type argument of `unary *' fts.c: In function `fts_build': fts.c:825: error: invalid type argument of `unary *' fts.c:849: error: invalid type argument of `unary *' fts.c: In function `fts_stat': fts.c:987: error: invalid type argument of `unary *' fts.c: In function `fts_palloc': fts.c:1152: error: invalid type argument of `unary *' fts.c: In function `fts_safe_changedir': fts.c:1227: error: invalid type argument of `unary *' fts.c:1238: error: invalid type argument of `unary *' make[2]: *** [fts.lo] Error 1 make[2]: Leaving directory `/tmp/openpkg- 2.4.2/rpm-4.2.1/rpmio' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/tmp/openpkg-2.4.2/rpm-4.2.1' make: *** [all] Error 2 + exit 2 runscript:ERROR: script returned non-null value snip This is very similar to what happened to me on AIX. Ralph had to tweak something within the rpm code to get past this.Yes, although RPM is rather portable, it unfortunately has some very bad pieces of C code (especially in this fts.c!) which have #ifdef sectionsfor all particular platforms. So, for each platform (class) one has toadd code there. I really hate RPM 4.2 for this (and hope it is gone in RPM 4.4)... Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
gcc and fortran
using CURRENT I can't seem to get gcc (now 4.0.1) to build with fortran support. I have gmp and mpfr installed, which worked for gcc 4.1 before, but I receive the following error now: checking for correct version of gmp.h... no configure: error: GMP with MPFR support is required to build f95
sqlite readline problem
when building sqlite from CURRENT with the readline option on, it links to termcap. is libtermcap in an openpkg package?
openpkg and debian
I have setup a debian sarge box, and have bootstrapped openpkg in to the system. After installing apache, the openpkg init.d script starts apache as the non-privileged user and I can see apache from localhost. However, from external connections I get a connection refused error, almost like there is a firewall running. I have not setup any non-default apps, and my test machines are all on the same subnet. Is there something extra I need to do to hook openpkg in to my debian system, or am I fighting some default security setting in debian? thanks
openpkg CURRENT
Are there some not-so-current repositories of CURRENT somewhere? I would like to get a hold of 1.32 version of boost to compare with the 1.33 version (which doesn't seem to work correctly).
webmin package
I was happy to see that webmin is in current now. I am wondering why many modules seem to be left out (e.g. nfs exports, samba)? I assume it is because this package is young and possibly not configured correctly, but I think most of those modules are standard issue, so some work must have went in to removing them.
Re: gcc and fortran
I tried the most recent gcc (gcc-4.0.1-20050922) and fortran now works, thanks! On 9/21/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Wed, Sep 21, 2005, Doug Henry wrote: using CURRENT I can't seem to get gcc (now 4.0.1) to build with fortran support.I have gmp and mpfr installed, which worked for gcc 4.1 before, but I receive the following error now: checking for correct version of gmp.h... no configure: error: GMP with MPFR support is required to build f95Ah, I think I've found it. Please retry with the latest gcc RPM from CURRENT. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
cvstrac question
I'm trying to run cvstrac under openpkg-2.5. there is a hint message that has suggested usage, which I tried. The database creation using cvstrac init command seems to have worked, but the http://localhost/openpkg-cgi/cvstrac/name/ address does not. I'm curious if anyone else has successfully used cvstrac?
Re: cvstrac question
nevermind.On 10/25/05, Doug Henry [EMAIL PROTECTED] wrote: I'm trying to run cvstrac under openpkg-2.5. there is a hint message that has suggested usage, which I tried. The database creation using cvstrac init command seems to have worked, but the http://localhost/openpkg-cgi/cvstrac/name/ address does not. I'm curious if anyone else has successfully used cvstrac?
cvs daemon question
there is a line in the rc.cvs file: cvs_daemon_listen=127.0.0.1:2401 what should this line be to allow for external connections. I tried just using the port number, and I tried netmask-style 192.168.1.0 and 192.168.1.255 and those didn't work. Is there some documentation on the options in this file? thanks, doug
Imake.tmpl
FYI, the Imake.tmpl (required for building xpm) under ubuntu, and I assume newer debian systems, is located in /etc/X11/config/cf, not the location pulled from the x11 package information used now. Not sure what the fix would be, but setting x11config to the previously mentioned location allows you to move on with your builds. -later
glark spec error
The spec file needs the following changed (prefix substitution is wrong). I think this error has existed for quite some time. %build %{l_shtool} subst \ -e 's;#!/bin/sh;#!%{l_prefix}/bin/ruby;' \ -e 's;exec ruby;#exec ruby;' \ glark
Re: gimp (2.5.0) build error
the gimp package from current works. On 11/2/05, Doug Henry [EMAIL PROTECTED] wrote: just wondering if anyone has seen these errors before. I get them when compiling gimp from opkg 2.5. if /mnt/tools/bin/cc -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -DHAVE_CONFIG_H -I. -I. -I../.. -I../.. -I../../app -I/mnt/tools/include/glib2 -I/mnt/tools/lib/glib2/include -I/mnt/tools/include -DG_LOG_DOMAIN=\Gimp-Composite\ -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -DGIMP_DISABLE_DEPRECATED -DGDK_MULTIHEAD_SAFE -DGTK_MULTIHEAD_SAFE -O2 -pipe -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -Wall -MT gimp-composite-mmx.o -MD -MP -MF .deps/gimp-composite-mmx.Tpo \ -c -o gimp-composite-mmx.o `test -f 'gimp-composite-mmx.c' || echo './'`gimp-composite-mmx.c; \ then mv -f .deps/gimp-composite-mmx.Tpo .deps/gimp-composite-mmx.Po; \ else rm -f .deps/gimp-composite-mmx.Tpo; exit 1; \ fi gimp-composite-mmx.c:835: warning: 'mmx_op_overlay' defined but not used gimp-composite-mmx.c: In function 'gimp_composite_addition_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:94: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_burn_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:149: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm5' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm5' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_darken_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:266: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_difference_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:300: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_grain_extract_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:556: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:556: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:556: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:565: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:565: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:565: error: unknown register name
Re: openpkg 2.5 IRIX
I put up a page on my wiki to use as a report of my efforts under IRIX. I have done many of the easy packages, and am now moving on to the ones which require patching. My motives/usage are different than that of the openpkg project, but maybe some of the info will be of use. Feel free to leave comments or send me email. http://www.brilligent.com/wikka.php?wakka=OpenPKGIRIX On 10/31/05, Matthias Kurz [EMAIL PROTECTED] wrote: On Mon, Oct 31, 2005 at 10:10:30AM -0500, Doug Henry wrote: I noticed that the -lgen change made it into 2.5, thanks! Is this the best place to post patches for IRIX? I am going to be rebuilding my toolchain, and I will post diffs against 2.5 if that is useful. If there is a wiki page I can post stuff there as well.Post your patches here (or better on openpkg-dev - i think that mostpeople are subscribed to both lists, anyway, though). I'm going to work as moderator. Thanks for your help ! (mk)--Matthias Kurz; Fuldastr. 3; D-28199 Bremen; VOICE +49 421 53 600 47 Im pr�motorischen Cortex kann jeder ein Held sein. (bdw) __The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: gimp (2.5.0) build error
The gimp (2.2.8) included with openpkg 2.5 can not be built with gcc4. There was a fix applied in the 2.2.9 release. On 11/2/05, Doug Henry [EMAIL PROTECTED] wrote: the gimp package from current works. On 11/2/05, Doug Henry [EMAIL PROTECTED] wrote: just wondering if anyone has seen these errors before. I get them when compiling gimp from opkg 2.5. if /mnt/tools/bin/cc -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -DHAVE_CONFIG_H -I. -I. -I../.. -I../.. -I../../app -I/mnt/tools/include/glib2 -I/mnt/tools/lib/glib2/include -I/mnt/tools/include -DG_LOG_DOMAIN=\Gimp-Composite\ -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -DGIMP_DISABLE_DEPRECATED -DGDK_MULTIHEAD_SAFE -DGTK_MULTIHEAD_SAFE -O2 -pipe -I/mnt/tools/include -I/mnt/tools/include/tiff -I/mnt/tools/include/gtk2 -I/mnt/tools/include/glib2 -Wall -MT gimp-composite-mmx.o -MD -MP -MF .deps/gimp-composite-mmx.Tpo \ -c -o gimp-composite-mmx.o `test -f 'gimp-composite-mmx.c' || echo './'`gimp-composite-mmx.c; \ then mv -f .deps/gimp-composite-mmx.Tpo .deps/gimp-composite-mmx.Po; \ else rm -f .deps/gimp-composite-mmx.Tpo; exit 1; \ fi gimp-composite-mmx.c:835: warning: 'mmx_op_overlay' defined but not used gimp-composite-mmx.c: In function 'gimp_composite_addition_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:94: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:101: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:121: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_burn_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:149: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm5' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:149: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm5' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:203: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_darken_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:266: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:266: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:280: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_difference_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:300: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:304: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm4' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm3' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm2' in 'asm' gimp-composite-mmx.c:327: error: unknown register name '%mm1' in 'asm' gimp-composite-mmx.c: In function 'gimp_composite_grain_extract_rgba8_rgba8_rgba8_mmx': gimp-composite-mmx.c:556: error: unknown register name '%mm6' in 'asm' gimp-composite-mmx.c:556: error: unknown register name '%mm7' in 'asm' gimp-composite-mmx.c:556: error: unknown register name '%mm0' in 'asm' gimp-composite-mmx.c:565: error
Re: gimp (2.5.0) build error
haven't tried that yet. there may be something more fundamentally wrong as well. gimp 2.2.9 from current builds, but produces a segfault. I also tweaked the 2.2.8 package to use gcc3, and that produces a segfault as well. I did a backtrace and I'm not sure it produces any useful info. #0 0xb723a103 in strlen () from /lib/tls/i686/cmov/libc.so.6 #1 0xb7c9743a in set_wm_name () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #2 0xb7c9940b in gdk_window_set_title () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #3 0xb7c998aa in gdk_window_new () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #4 0xb7ca0adb in gdk_display_open () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #5 0xb7c7380d in gdk_display_open_default_libgtk_only () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #6 0x08306f57 in gtk_init_check () #7 0x0807fde8 in ?? () #8 0xbfc818d0 in ?? () #9 0xbfc818d4 in ?? () #10 0x0007 in ?? () #11 0x0836923c in _IO_stdin_used () #12 0xb7e8019c in libintl_nl_default_dirname () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #13 0xbfc818e0 in ?? () #14 0xbfc81778 in ?? () #15 0xb7b7eeec in libintl_textdomain () from /mnt/tools/lib/gimp/libgimpwidgets-2.0.so.0 #16 0x0807e369 in ?? () #17 0xbfc818d0 in ?? () #18 0xbfc818d4 in ?? () #19 0xbfc818d4 in ?? () #20 0x2e32 in ?? () #21 0x32312e36 in ?? () #22 0x6b2d392d in ?? () #23 0xb765bf78 in ?? () #24 0xb765bf68 in ?? () #25 0x08072051 in ?? () #26 0xbfc817c0 in ?? () #27 0xb7f7c060 in _dl_debug_state () from /lib/ld-linux.so.2 #28 0xb71e4ea2 in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6 #29 0x0807d991 in ?? () On 11/3/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Thu, Nov 03, 2005, Doug Henry wrote: The gimp (2.2.8) included with openpkg 2.5 can not be built with gcc4.There was a fix applied in the 2.2.9 release.The problem just is that the code difference between Gimp 2.2.8 and2.2.9 is 8.3 MB diff(3) output and the change to gimp-composite-mmx.cdoesn't look sufficient enough. Are you able to extract the fix for GCC4? Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
openpkg 2.5 mysql
I haven't been able to get mysql under openpkg 2.5 to work (debian). I always get an error that the socket file could not be created. If I build the mysql package from 2.4, all is well.
Re: openpkg 2.5 mysql
I'll give it a try...thanks. On 11/7/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Mon, Nov 07, 2005, Doug Henry wrote: I haven't been able to get mysql under openpkg 2.5 to work (debian).I always get an error that the socket file could not be created.If I build the mysql package from 2.4, all is well.Yes, we already identified this and tried to fix it in OpenPKG-CURRENT.I've now backported the fix to OpenPKG 2.5. Please retry withmysql-4.1.14-2.5.1 now. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: openpkg 2.5 pcre configure error
debian stable (3.1). I have an ubuntu breezy install that works correctly. On 11/16/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Wed, Nov 16, 2005, Doug Henry wrote: I get this error when building pcre (2.5) under debian: checking build system type... Invalid configuration `i686-pc-linux-': machine `i686-pc-linux' not recognized configure: error: /bin/sh ./config.sub i686-pc-linux- failed When I run ./config.guess I get i686-pc-linux-, but I believe I should get i686-pc-linux-gnu.Not sure if anyone has seen this problem before. On what particular Linux flavor is this? Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: openpkg 2.5 pcre configure error
alright, I'll check it out. I have a sarge install: Linux natasha 2.6.8-2-k7 #1 Thu May 19 18:03:29 JST 2005 i686 GNU/Linux I'll poke around in the config.guess script and find out where the missing gnu is suppose to come from. -thanks On 11/17/05, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Wed, Nov 16, 2005, Doug Henry wrote: debian stable (3.1).I have an ubuntu breezy install that works correctly. [...]Then something has to be locally broken on your particular Debian box. We've also an up-to-date Debian 3.1 box in the OpenPKG Foundation andthere the pcre package builds just fine. The outputs are:| [EMAIL PROTECTED]:/tmp/rse/openpkg/pcre-6.4 [E=/openpkg-dev T=/openpkg-dev S=src] | developer$ uname -a| Linux rm2.openpkg.net 2.6.13.2 #1 SMP Tue Sep 20 14:39:59 CEST 2005 i686 GNU/Linux| [EMAIL PROTECTED]:/tmp/rse/openpkg/pcre- 6.4 [E=/openpkg-dev T=/openpkg-dev S=src]| developer$ ./config.guess| i686-pc-linux-gnu| developer$ Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
graphviz current
graphviz package in current seems broken. none of the plugins work (e.g. layout plugin), I think you should be able to run dot -c and generate the prefix/lib/graphviz/config file, but that doesn't work either. I have not tried the 2.5 version to see if it is also broke. My litmus test has been using doxygen to generate documentation, the dot graph generation always fails. Usually the maintainers are very fast with these problems, if I get around to comparing with old graphviz packages (it worked in the past), I will try to post a fix. -later
boost build issues
I am unable to build the current boost (1.33.0) and was going to fix it. does anyone know of a log file that boost generates? I get the following: FileClone stage/lib/libboost_program_options-mt.so FileClone stage/lib/libboost_program_options-mt.so.1.33.0 ...failed updating 6 targets... ...skipped 8 targets... ...updated 738 targets... error: Bad exit status from /mnt/tools/RPM/TMP/rpm-tmp.29282 (%build) I am unable to determine what happened during the failed updating 6 targets. -thanks
openpkg.sh problems
from openpkg-2.5.1 forward I can no longer bootstrap from the openpkg.src.sh file. I receive the following messages with 2.5.1 and the 0601 version in current. I can correctly bootstrap from openpkg-2.5.0.src.sh and then upgrade to 2.5.1. I am running ubuntu (breezy). [EMAIL PROTECTED]:~/tmp$ sh /export/share/openpkg/devel.testing/openpkg-20060101-20060101.src.sh --prefix=/mnt/tools-testing --tag=testing : command not foundkg/devel.testing/openpkg-20060101-20060101.src.sh: line 25: : command not foundkg/devel.testing/openpkg-20060101-20060101.src.sh: line 35: ': not a valid identifierel.testing/openpkg-20060101-20060101.src.sh: line 39: export: `LC_CTYPE : octal number out of range.testing/openpkg-20060101-20060101.src.sh: line 40: umask: 022 : command not foundkg/devel.testing/openpkg-20060101-20060101.src.sh: line 41: /export/share/openpkg/devel.testing/openpkg-20060101-20060101.src.sh: line 44: syntax error near unexpected toke' `do 'export/share/openpkg/devel.testing/openpkg-20060101-20060101.src.sh: line 44: `do [EMAIL PROTECTED]:~/tmp$
samba swat
how do I enable swat support in samba. I have compiled the package with swat support, and the executable is in sbin, but obviously the normal inet config does not apply. -thanks
Re: samba swat
thanks, I'm using this as my /opkg/etc/xinetd.d/swat file, seems to give me access: service swat { port = 901 socket_type = stream wait = no #> user = root server = /opkg/sbin/swat log_on_failure += USERID #disable = yes } On 1/18/06, Bill Campbell [EMAIL PROTECTED] wrote: On Wed, Jan 18, 2006, Doug Henry wrote: I guess it can, I assumed (my fault) there would be a config supplied in xinetd.d.I have a slightly hacked version of Release 2.3 SAMBA that sets this up when the with_swat option is turned on.ftp://ftp.celestial.com/private/ftp.openpkg.org/release/2.3/SRC/samba-3.0.11-20050901.src.rpm On 1/18/06, Bill Campbell [1]bill@celestial.com wrote: On Wed, Jan 18, 2006, Doug Henry wrote: how do I enable swat support in samba.I have compiled the package with swat support, and the executable is in sbin, but obviously the normal inet config does not apply. Why not? You will have to configure either inetd or xinetd. Bill -- INTERNET: [EMAIL PROTECTED]Bill Campbell; Celestial Software LLC URL: [3]http://www.celestial.com/PO Box 820; 6641 E. Mercer Way FAX:(206) 232-9186Mercer Island, WA 98040-0820; (206) 236-1676 Intaxication: Euphoria at getting a refund from the IRS, which lasts until you realize it was your money to start with. ___ ___ The OpenPKG Project[4]www.openpkg.org User Communication List[5]openpkg-[EMAIL PROTECTED]References 1. mailto:[EMAIL PROTECTED] 2. mailto: [EMAIL PROTECTED] 3. http://www.celestial.com/ 4. http://www.openpkg.org/ 5. mailto: openpkg-users@openpkg.org--Bill--INTERNET: [EMAIL PROTECTED]Bill Campbell; Celestial Software LLCURL: http://www.celestial.com/PO Box 820; 6641 E. Mercer WayFAX:(206) 232-9186Mercer Island, WA 98040-0820; (206) 236-1676``A human being should be able to change a diaper, plan an invasion,butcher a hog, conn a ship, design a building, write a sonnet, balanceaccounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitchmanure, program a computer, cook a tasty meal, fight efficiently, diegallantly.Specialization is for insects.'' Robert Heinlein__ The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: FSL on IRIX 6.5
FYI, I built fsl-2.5.0 on my SGI and it builds just fine. I know mine works is not very helpful, I will try to use some time to compare my inet stuff to see if there are differences. I'm using gcc-3.4.4 (from 2.4 release), which is about the only thing I can think of that could affect the compile other than the OS itself. I would stay away from gcc4/binutils at all costs on IRIX for a while. I have had VERY strange things happen in my development work, gcc3 runs like a champ. Specs: IRIX 6.5.16 on IP30/R12k processors. On 1/23/06, Doug Henry [EMAIL PROTECTED] wrote: you could probably add the define (-DINET6 or whatever) to the CPPFLAGS line in the spec file. If that doesn't work use CFLAGS and/or CXXFLAGS. Watch the compile to make sure the configure process used your flag (you should see it in the compile line). On 1/20/06, Doug Summers [EMAIL PROTECTED] wrote: Ralf S. Engelschall wrote: On Thu, Jan 19, 2006, Doug Summers wrote:Ralf S. Engelschall wrote:On Thu, Jan 19, 2006, Doug Summers wrote: You can't really disable it explicitly as the contained OSSP salibrary's source code just detects IPV6 via some #ifdefs on the symbolAF_INET6. But in your case the problem is that the #include is missing for the struct sockaddr_in6. Usually this has to be provided bynetinet/in.h but on your platform this isn't the case. Can you performa grep -r sockaddr_in6 /usr/include/ and give us the output, please? grep -r sockaddr_in6 /usr/include//usr/include/net/raw_cb.h:#include netinet/in.h /* for sockaddr_in6 *//usr/include/net/raw_cb.h:structsockaddr_in6 raw_dst6;/*dst address for rawintr *//usr/include/net/raw_cb.h:structsockaddr_in6 raw_src6;/*src address for sbappendaddr *//usr/include/netinet/in_pcb.h:structsockaddr_in6 spare_dst;/usr/include/netinet/in.h:struct sockaddr_in6 { /usr/include/netinet/in.h:#define satosin6(sa)((struct sockaddr_in6*)(sa))/usr/include/rapi_lib.h:struct sockaddr_in6 sender; Hmmm... interesting, the stuff actually is in netinet/in.h. But Iguess there is a nice #ifdef somewhere around it which prevents it frombeing activated. Can you send us the complete /usr/include/netinet/in.h file, too?[...]#ifdef INET6/* * IPv6 socket address */#define SIN6_LENstruct sockaddr_in6 { uint8_tsin6_len;u_int8_tsin6_family;/* XXX6 should be sa_family_t but it is a short */in_port_t sin6_port;uint32_t sin6_flowinfo;structin6_addr sin6_addr; };[...] Ah, there it is! Hmmm... INET6?! Who should set this define? I guess the IRIX guys think the application has to set this. Well, you can then try to define this with something like -DINET to workaround your problems. Ralf S. Engelschall[EMAIL PROTECTED] www.engelschall.com Sorry, Ralph - I'm not a developer. How would I do this???Doug__The OpenPKG Projectwww.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: NFS question
The system chmod. I can't remember if I have tried the openpkg (coreutils) one.On 2/24/06, David M. Fetter [EMAIL PROTECTED] wrote:Are you using the IRIX chmod or the gnu chmod from within OpenPKG? On Fri, 2006-02-24 at 10:16 -0500, Doug Henry wrote: I have an NFS issue that I wanted to see if anyone has seen before. My basic setup is a linux server, and mixed client machines.On the linux server there is a large raid which is a giant playground of data and has the following share defined: /raid/data*(sync,rw,all_squash,anonuid=1,anongid=1) I do this because ownership and permissions are meaningless on this data share, and it keeps all data files usable by all...all of the time. It works perfectly under linux, solaris, and through samba to my windows clients.On my IRIX machine (stay with me...it may not be an irix issue) I can create and delete files properly, but if I do a chmod, I get an operation denied message.Strangely, irix seems to check if I am the owner of the file before performing the chmod and fails, but all other basic file operations work as expected. Any and all suggestions welcome. -thanks--David M. Fetter - UNIX System AdministratorChange is inevitable. Growth is optional. ~unknown-BEGIN PGP SIGNATURE- Version: GnuPG v1.4.2 (GNU/Linux)iD8DBQBD/zI+/aYAl/wcnokRAjbjAJ0dxH7TREjkAnNCP7cgUllKaMl5DACgkF/olUZA8sJfjtdoab47m2ynJi4==/iX0-END PGP SIGNATURE-
coreutils problem report
FYI problem report:when building coreutils::with_legacy=yes from current, I receive the following build error:++ echo gyes++ sed -e 's;^g;;'+ ln -s gyes yesln: `yes': File exists+ exit 1 error: Bad exit status from /mnt/tools-testing/RPM/TMP/rpm-tmp.44656 (%install)
registry password reset/recovery
I was curious if there is a plan to allow for registered openpkg accounts to reset their passwords (just in case we would forget them:)).-thanks
fontconfig error
the fontconfig package in current fails during build with the following error: 'uintptr_t' undeclaredI added #include stdint.h to src/fcint.h and it fixes the problem.-later
Re: Is there a tool for handling package dependencies for OpenPKG
I don't have anything fancy, but I use packages from many of the openpkg releases to create my own set of tools. For example, I might use gcc from 2.4 release, many things from 2.5 release and some misc stuff from current. I have some cpp code that reads a list of packages and pulls the xml chunk from the 00INDEX.rdf file and adds it to a new xml tree. Basically works like an xml filter, allowing me to combine the 00INDEX files to create a custom release structure. I can package up my filter code if it sounds like something you might be able to use. Adding my own rpms if pretty easy, I just maintain my own 00INDEX.rdf file with the info for my custom packages and modify the top-level 00INDEX file to point to it. You may want to bzip2 -d 00INDEX.rdf.bz2 and see what they look like. The 00INDEX.rdf files don't have to be compressed to use, so I usually just leave them as xml so I can tweak easily. -laterOn 3/24/06, Olaf Mersmann [EMAIL PROTECTED] wrote: * Ralf S. Engelschall [EMAIL PROTECTED] [060322 08:19]: On Wed, Mar 22, 2006, Dietrich Bollmann wrote: Is there some installation management tool like 'yum' for dealing with rpm package dependencies? Yes, install the openpkg-tools package. It provides an openpkg build tool which is able to fetch the RDF index files from the FTP service and automatically resolve dependencies. Run openpkg man build for details after installing openpkg-tools.`openpkg build` works well on a system which actually build sourcepackages, but I've find it annoying, that it will install (and uninstall) build dependencies even if a current binary package is inthe RPM repository. Is there a way to change this behavior or am Idoing something seriously wrong?I'd also love to see a tutorial on the 'right' way to build custom source repositories. Specifically I'd like to know the 'proper' way tocreate an overlay to the official source repository.On another note: I noticed both yum and urpmi are in -current butneither would build out of the box for me on Debian 3.1 or Solaris10. Is anyone using these? Is there any interest in getting them towork within the OpenPKG framework?-- Olaf__The OpenPKG Project www.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
installing a fake package
Is there a mechanism to fake the installation of a package? For example, under IRIX the perl package will not compile (as provided) using gcc. The system perl seems to work just fine for the packages I have tried requiring perl, so the ability to make openpkg think it has perl installed would be very helpful. -thanks
cmake spec file tweak
I noticed cmake has made it in to current. I attached the cmake.spec file I have been using for quite a while to possibly help with future cmake packages. The current package does not build the very useful curses front-end (ccmake). Obviously making ncurses a build requirement is useful here, but there are some cmake specific changes that need to be made as well. CMake builds in two steps: bootstrap and final. During bootstrap the cmake program is built, and during final, cmake is used to configure and build the entire package. In order to get options to the final step, they need to be passed via the --init option. Again, I don't claim to be a great packager, but I hope some of this info will prove useful. cmake.spec Description: Binary data
Re: installing a fake package
So it looks like what I am doing now (removing the build and install stuff from the spec file) is the way to do it. I thought standard rpm had an inject option which would let you put a package in the database without actually having a package, but I haven't been able to find that option under openpkg. -thanksOn 8/28/06, Doug Summers [EMAIL PROTECTED] wrote: Doug Henry wrote: Is there a mechanism to fake the installation of a package?For example, under IRIX the perl package will not compile (as provided) using gcc.The system perl seems to work just fine for the packages I have tried requiring perl, so the ability to make openpkg think it has perl installed would be very helpful. -thanksIn the past I did this on AIX by hacking the .spec and removing all of the build/install entries, basically creating an empty package. Didn'thelp much when it was time to upgrade though.Doug__The OpenPKG Project www.openpkg.orgUser Communication Listopenpkg-users@openpkg.org
Re: installing a fake package
I guess justdb is what I would want. I don't really care that I need the package (I have it for other platforms anyway), but modifying it just so I can not install it is strange.On 8/30/06, Tres Seaver [EMAIL PROTECTED] wrote: Doug Henry wrote: So it looks like what I am doing now (removing the build and install stuff from the spec file) is the way to do it.I thought standard rpm had an inject option which would let you put a package in the database without actually having a package, but I haven't been able to find that option under openpkg.Stock RPM has'--justdb':$ rpm --help | grep -C 1 just-i, --installinstall package(s) --justdb update the database, but do notmodify the filesystemTres.--=== Tres Seaver+1 202-558-7113[EMAIL PROTECTED]Palladion Software Excellence by Designhttp://palladion.com __The OpenPKG Projectwww.openpkg.orgUser Communication List openpkg-users@openpkg.org
PHP UTF-8 support
I need php with utf8_encode and utf8_decode support. I was wondering which options in the apache package will turn that on for me. My first guess is apache::with_mod_php_iconv, but I don't want to do a bunch of rebuilding just to find out. -thanks
Re: PHP UTF-8 support
-Dapache::with_mod_php_iconv=yes-Dapache::with_mod_php_mbstring=yesOn 11/6/06, Doug Henry [EMAIL PROTECTED] wrote:I need php with utf8_encode and utf8_decode support. I was wondering which options in the apache package will turn that on for me. My first guess is apache::with_mod_php_iconv, but I don't want to do a bunch of rebuilding just to find out. -thanks
binutils 2.17 fails to build
The binutils 2.17 package in stable and current fails to build on many of my systems. The errors below are very common across my systems. /local/devel/bin/cc -DHAVE_CONFIG_H -I. -I.././binutils -I. -D_GNU_SOURCE -I. -I.././binutils -I../bfd -I.././binutils/../bfd -I.././binutils/../include -I.././binutils/../intl -I../intl -DLOCALEDIR=\/local/devel/share/locale\ -Dbin_dummy_emulation=bin_vanilla_emulation -W -Wall -Wstrict-prototypes -Wmissing-prototypes -Werror -pipe -O3 -mtune=pentium4 -DOPENPKG_OS_LINUX -c not-strip.c /local/devel/bin/cc -DHAVE_CONFIG_H -I. -I.././binutils -I. -D_GNU_SOURCE -I. -I.././binutils -I../bfd -I.././binutils/../bfd -I.././binutils/../include -I.././binutils/../intl -I../intl -DLOCALEDIR=\/local/devel/share/locale\ -Dbin_dummy_emulation=bin_vanilla_emulation -W -Wall -Wstrict-prototypes -Wmissing-prototypes -Werror -pipe -O3 -mtune=pentium4 -DOPENPKG_OS_LINUX -c wrstabs.c cc1: warnings being treated as errors wrstabs.c: In function 'stab_tag_type': wrstabs.c:1297: warning: 'size' is used uninitialized in this function make[4]: *** [wrstabs.o] Error 1 make[3]: *** [all-recursive] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-binutils] Error 2 make: *** [all] Error 2 error: Bad exit status from /local/devel/RPM/TMP/rpm-tmp.68586 (%build)
Re: binutils 2.17 fails to build
good build from current...thanks! On 11/17/06, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Fri, Nov 17, 2006, Doug Henry wrote: The binutils 2.17 package in stable and current fails to build on many of my systems. The errors below are very common across my systems. /local/devel/bin/cc -DHAVE_CONFIG_H -I. -I.././binutils -I. -D_GNU_SOURCE -I. -I.././binutils -I../bfd -I.././binutils/../bfd -I.././binutils/../include -I.. /./binutils/../intl -I../intl -DLOCALEDIR=\/local/devel/share/locale\ -Dbin_dummy_emulation=bin_vanilla_emulation -W -Wall -Wstrict-prototypes -Wmissing-prototypes -Werror -pipe -O3 -mtune=pentium4 -DOPENPKG_OS_LINUX -c not-strip.c /local/devel/bin/cc -DHAVE_CONFIG_H -I. -I.././binutils -I. -D_GNU_SOURCE -I. -I.././binutils -I../bfd -I.././binutils/../bfd -I.././binutils/../include -I.. /./binutils/../intl -I../intl -DLOCALEDIR=\/local/devel/share/locale\ -Dbin_dummy_emulation=bin_vanilla_emulation -W -Wall -Wstrict-prototypes -Wmissing-prototypes -Werror -pipe -O3 -mtune=pentium4 -DOPENPKG_OS_LINUX -c wrstabs.c cc1: warnings being treated as errors wrstabs.c: In function 'stab_tag_type': wrstabs.c:1297: warning: 'size' is used uninitialized in this function make[4]: *** [wrstabs.o] Error 1 make[3]: *** [all-recursive] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-binutils] Error 2 make: *** [all] Error 2 error: Bad exit status from /local/devel/RPM/TMP/rpm-tmp.68586 (%build) Although I've no clue why this warnings occurs just for you (actually I've looked into the 2.17 source and do not understand why it occurs at all or not on all platforms), I've disabled this -Werror stuff for OpenPKG now as we are packaging the software and not developing it. Try the latest binutils from CURRENT and 2-STABLE. The problem should be now gone. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ The OpenPKG Projectwww.openpkg.org User Communication List openpkg-users@openpkg.org
options description
I was wondering if the options to packages are documented anywhere? Just a simple description of what they provide is sufficient. Many are self explanatory, but things like imapd::with_replication are not, at least not to me. -thanks
gcc 32/64
I am trying to compile openpkg stable 1018 snapshot on a 64-bit machine. Once the 32-bit stuff starts building I get the attached error. I get the same error if I build the gcc from current using the multilib option. Without multilib everything works fine. I can build 32-bit using the system compiler and the -m32 flag. Anyone have any ideas? LIBGCC2_CFLAGS=-O2 -O2 -O -DIN_GCC-W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED -m32 \ MULTILIB_CFLAGS= -m32 T=32/ 32/crtbegin.o 32/crtend.o 32/crtbeginS.o 32/crtendS.o 32/crtbeginT.o 32/crtfastmath.o ./xgcc -B./ -B/tools/x86_64-unknown-linux-gnu/bin/ -isystem /tools/x86_64-unknown-linux-gnu/include -isystem /tools/x86_64-unknown-linux-gnu/sys-include -L/tools/RPM/TMP/gcc-4.1.1/obj/gcc/../ld -O2 -O2 -O -DIN_GCC-W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -isystem ./include -I. -I32 -I/tools/RPM/TMP/gcc-4.1.1/obj/../gcc -I/tools/RPM/TMP/gcc-4.1.1/obj/../gcc/32 -I/tools/RPM/TMP/gcc-4.1.1/obj/../gcc/../include -I/tools/RPM/TMP/gcc-4.1.1/obj/../gcc/../libcpp/include -m32 -g0 -finhibit-size-directive -fno-inline-functions -fno-exceptions -fno-zero-initialized-in-bss -fno-unit-at-a-time -fno-omit-frame-pointer -fno-asynchronous-unwind-tables \ -c /tools/RPM/TMP/gcc-4.1.1/obj/../gcc/crtstuff.c -DCRT_BEGIN \ -o 32/crtbegin.o In file included from /usr/include/features.h:346, from /usr/include/stdio.h:28, from /tools/RPM/TMP/gcc-4.1.1/obj/../gcc/tsystem.h:90, from /tools/RPM/TMP/gcc-4.1.1/obj/../gcc/crtstuff.c:68: /usr/include/gnu/stubs.h:7:27: error: gnu/stubs-32.h: No such file or directory make[4]: *** [32/crtbegin.o] Error 1 make[3]: *** [extra32] Error 2 make[2]: *** [stmp-multilib] Error 2 make[1]: *** [stage1_build] Error 2 make: *** [bootstrap-lean] Error 2 error: Bad exit status from /tools/RPM/TMP/rpm-tmp.78133 (%build) RPM build errors: Bad exit status from /tools/RPM/TMP/rpm-tmp.78133 (%build)
strange apache echo build error
I am getting the attached error when building apache (1018). It appears there is some sort of echo problem during apache configure. Just wondering if anyone has worked around this error yet. typescript Description: Binary data
Re: strange apache echo build error
This has something to do with ubuntu edgy, dapper seems to work as expected. On 12/4/06, Doug Henry [EMAIL PROTECTED] wrote: I am getting the attached error when building apache (1018). It appears there is some sort of echo problem during apache configure. Just wondering if anyone has worked around this error yet.
jabberd configuration
I finally got around to setting up a jabber server. I am using the jabberd package included with openpkg (stable). Reading the jabberd docs and looking at the default configuration files, it seems the server should work for localhost out of the box. However, all of the processes shutdown. The log files state that the sm program terminated, causing all others to exit as well. Are there any jabberd users out there that could offer some insight? -thanks
patch needed for swig or python
Not sure which is broken yet, but there is an incompatibility between swig and python in openpkg when using C++. I haven't found the patch yet, but debian swig/python works. The process I use is: g++ -c file.cpp swig -c++ -python file.i g++ -c `python-config --includes` file_wrap.cxx during the compile of the wrapper, I receive the following error always. error: invalid conversion from 'const char*' to 'char*'
Re: patch needed for swig or python
Built a swig-1.3.31 package and confirmed generated swig code works with python 2.5. On 12/19/06, Doug Henry [EMAIL PROTECTED] wrote: I found some info that suggests versions of swig prior to 1.3.30 are incompatible with python 2.5. On 12/19/06, Doug Henry [EMAIL PROTECTED] wrote: Not sure which is broken yet, but there is an incompatibility between swig and python in openpkg when using C++. I haven't found the patch yet, but debian swig/python works. The process I use is: g++ -c file.cpp swig -c++ -python file.i g++ -c `python-config --includes` file_wrap.cxx during the compile of the wrapper, I receive the following error always. error: invalid conversion from 'const char*' to 'char*'
Re: patch needed for swig or python
I think my last post clears up the issue. An interface change was made in python 2.5 that broke swig prior to 1.3.30. So I think its simply a problem that openpkg includes python 2.5 and swig 1.3.29. I downloaded swig 1.3.31, updated the swig.spec file, build the package and everything works. You most likely run in to this sort of version creep problem in current all the time, as it is a fast moving target. For the stable branch, it would be worth considering an update to swig for the sake of inter-package compatibility. -thanks Well, I guess Debian's SWIG is a different version than OpenPKG's SWIG and OpenPKG's SWIG seems to generate C++ code which our GCC dislikes. Look at the line GCC complains about: the generated code certainly has a difference there (or at the place of a variable declaration in case the line GCC complains about uses a variable). But in any way, I think you have to report this to either the Python and/or the SWIG guys, as only those will be able to really help... Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ The OpenPKG Projectwww.openpkg.org User Communication List openpkg-users@openpkg.org
sqlite spec error
The last two or three sqlite updates (from current) have failed with the same error. I believe it is related to the addition of the odbc option. The error I receive after update is: Preparing...### [100%] 1:sqlite ### [100%] /TMP/rpm-tmp.37738: line 3: syntax error near unexpected token `fi' /TMP/rpm-tmp.37738: line 3: `fi' error: %preun(sqlite-3.3.8-20061218) scriptlet failed, exit status 2 I believe the software updates, but the database must not get updated, because a build -Ua will always rebuild sqlite and fail.
texinfo build error
The texinfo package (stable and current) will fail to build if curses library is not installed on the host system. Installing ncurses from openpkg does not fix the problem. Adding %{l_cppflags ncurses} to CFLAGS seems to help.
Re: texinfo build error
I think my previous message bounced. I added the following two lines to the build for texinfo, which seems to fix the build issues. CPPFLAGS=%{l_cppflags ncurses} \ LDFLAGS=%{l_ldflags} \ On 1/25/07, Doug Henry [EMAIL PROTECTED] wrote: The texinfo package (stable and current) will fail to build if curses library is not installed on the host system. Installing ncurses from openpkg does not fix the problem. Adding %{l_cppflags ncurses} to CFLAGS seems to help.
openpkg on amd64
I have been building many openpkg packages under linux amd64. Most packages work fine, but many produce errors like the following: /local/tools/Linux/bin/cc -pipe -O3 -fPIC -shared -fPIC -I../src -I/local/tools/Linux/include pcre-module.c -o pcre-module.so-L/local/tools/Linux/lib -lpcre -Wl,-R/local/tools/Linux/lib:/local/tools/Linux/lib:/local/tools/Linux/lib -lm /local/tools/Linux/bin/ld: /local/tools/Linux/lib/libpcre.a(pcre_compile.o): relocation R_X86_64_32S against `_pcre_utf8_table4' can not be used when making a shared object; recompile with -fPIC /local/tools/Linux/lib/libpcre.a: could not read symbols: Bad value collect2: ld returned 1 exit status make[1]: *** [pcre-module.so] Error 1 make: *** [all] Error 2 Has anyone conquered this problem already? I have been googling, but there doesn't seem to be a clear and concise explanation of this error. -doug
Re: [2.20061018 stable] gcc with_fortran fails to build
gcc with fortran from current (4.1.2) builds for me under debian/ubuntu. If you haven't already, I would build gcc without fortran, and then rebuild it with fortran so it builds using the same version of openpkg gcc and not the system compiler. I have seen cases with several packages where the build works with the openpkg gcc and not the system gcc. If you are already doing this, it would have to be some sort of system header weirdness. -good luck On 2/20/07, David M. Fetter [EMAIL PROTECTED] wrote: Thanks. I'll try that out. On Tue, 2007-02-20 at 16:18 -0800, Bill Campbell wrote: On Tue, Feb 20, 2007, David M. Fetter wrote: Ya, I did that already. It has the same failure. It looks like some structure is missing members. Look at line 2440 in this file to see what structure is referenced with the term gfc_expr: /usr/local/RPM/USER/TMP/gcc-4.1.1/obj/../gcc/fortran/arith.c Then try to figure out which header file contains the structure and whether you can see gfc_expr anyplace. I frequently use the attached perl script to create a pre-processed tmp.c file so I can see exactly what's going on. Capture the output of the build so you have the compile line that failed, then replace the first ``gcc'' or ``cc'' part with this script redirecting the output to a file for analysis (watch out for -o options as the preprocessed output will end up there if it's present). The line with do.*csspath can be removed, and the spitshell stuff at the top fixed to fit your system. ... Bill -- INTERNET: [EMAIL PROTECTED] Bill Campbell; Celestial Software LLC URL: http://www.celestial.com/ PO Box 820; 6641 E. Mercer Way FAX:(206) 232-9186 Mercer Island, WA 98040-0820; (206) 236-1676 ``One of the common failings among honorable people is a failure to appreciate how thoroughly dishonorable some other people can be, and how dangerous it is to trust them.'' - Thomas Sowell -- David M. Fetter - Portland State University - UNIX Systems Administrator Reality is merely an illusion, albeit a very persistent one. ~Einstein
Re: openpkg build -r not work with local fs mirror
I take it relative paths are no longer supported? Previously I had the line -r . in my ~/.openpkg/build file. On 2/26/07, Thomas Lotterer [EMAIL PROTECTED] wrote: Ralf S. Engelschall[EMAIL PROTECTED] wrote on 2007-02-23 11:04: On Fri, Feb 23, 2007, Victor G. Bolshakov wrote: As i understand, openpkg-tools not work with local fs mirror from OpenPKG 2.20070221: Sure, it is. Just specify -r file:///path instead of -r /path on the openpkg build command line until Thomas adds a wordaround which does this automatically. I changed openpkg tools today [1] in the hope to create a behavior which is compatible and intuitive. It is now possible to use a URL or an absolute path pointing to a package repository. Using URL style, including file://, engages openpkg curl to download a local copy of the package before use while using an absolute path, beginning with /, direct access is used. $ openpkg build -r ftp://download.example.com/repo foo $ openpkg build -r http://download.example.com/repo foo $ openpkg build -r https://download.example.com/repo foo $ openpkg build -r file://download.example.com/repo foo ... openpkg curl -o foo.src.rpm ...//download.example.com/repo/foo.src.rpm openpkg rpm --rebuild foo.src.rpm openpkg rpm -Uvh foo.rpm rm foo.src.rpm # unless -k (keep) option specified or $ openpkg build -r /mirror/repo foo ... openpkg rpm --rebuild /mirror/foo.src.rpm openpkg rpm -Uvh foo.rpm I hope that makes sense to everyone. [1] http://cvs.openpkg.org/chngview?cn=34214 - openpkg-tools-1.4.2 - openpkg-20070226 -- http://thomas.lotterer.net __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
Re: openpkg build -r not work with local fs mirror
Absolute paths and ./ paths seem like very acceptable options. Thanks! On 3/2/07, Thomas Lotterer [EMAIL PROTECTED] wrote: Doug Henry [EMAIL PROTECTED] wrote on 2007-02-28 20:23: I take it relative paths are no longer supported? Previously I had the line -r . in my ~/.openpkg/build file. I don't wanna say relative paths are supported because the openpkg setUID wrapper may change directory under certain circumstances. One case I know of is that it changes to $HOME when $PWD becomes inaccessible, which can happen if an instance manager has the current working directory on a NFS drive with the usual root-nobody mapping and executes a command which requires root privileges. That said I also do not want to introduce artificial limitations. So I added support for relative paths identified by ./ (no plans to add .. or all but leading / detection). Available with openpkg-tools-1.4.4, included with openpkg-20070302. -- http://thomas.lotterer.net __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
bootstrap using sun compilers
Has anyone successfully bootstrapped openpkg using sun studio 11 or 12 compilers? I am getting an error building openssl (redefinition of ASN1_dup). Is it possible to define the compilers and options used during the bootstrap, sort of like with the .rpmmacros file? -later
Re: rsync
I would like to rsync the current/SRC branch of openpkg for offline use. I have not done this before and I am looking for information on how it is done for openpkg. On 7/15/07, Andy Harrison [EMAIL PROTECTED] wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 7/15/07, Doug Henry wrote: Does rsync still work? I am trying to sync the current branch using the following command but I always get a failure. rsync rsync://rsync.openpkg.org/openpkg-ftp/current/SRC/ First, it is important that you read this web page. How To Ask Questions The Smart Way [http://www.catb.org/~esr/faqs/smart-questions.html] Hopefully that will encourage you to actually share important information with us like the error message. - -- Andy Harrison public key: 0x67518262 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.6 (GNU/Linux) Comment: http://firegpg.tuxfamily.org iD8DBQFGmiI4NTm8fWdRgmIRAq7OAJ0dWdn3Ez+zjZffiusRJ1zV9MELzwCgoCCY BsC3Oo0IOJDhgD3oKEMga3s= =FjpC -END PGP SIGNATURE- __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
freetype build broken on OSX
I receive the following error when compiling freetype (current) on OSX. It seems to be related to the make setup command used. If I run 'openpkg rpm -bp freetype.spec' and then ./configure; make; make install from the TMP/freetype-2.3.5 folder it works as expected. If I run the 'make setup ...' command I receive the target install error message. ar -r objs/libfreetype.a ./objs/ftsystem.o ./objs/ftdebug.o ./objs/ftinit.o ./objs/ftbase.o ./objs/ftbbox.o ./objs/ftbdf.o ./objs/ftbitmap.o ./objs/ftglyph.o ./objs/ftgxval.o ./objs/ftmm.o ./objs/ftotval.o ./objs/ftpfr.o ./objs/ftstroke.o ./objs/ftsynth.o ./objs/fttype1.o ./objs/ftwinfnt.o ./objs/ftxf86.o ./objs/ftlcdfil.o ./objs/ftgasp.o ./objs/ftpatent.o ./objs/truetype.o ./objs/type1.o ./objs/cff.o ./objs/type1cid.o ./objs/pfr.o ./objs/type42.o ./objs/winfnt.o ./objs/pcf.o ./objs/bdf.o ./objs/sfnt.o ./objs/autofit.o ./objs/pshinter.o ./objs/raster.o ./objs/smooth.o ./objs/ftcache.o ./objs/ftgzip.o ./objs/ftlzw.o ./objs/psaux.o ./objs/psnames.o ar: creating archive objs/libfreetype.a + exit 0 Executing(%install): env -i /Users/doug/tools/lib/openpkg/bash --norc --noprofile --posix -e /Users/doug/tools/RPM/TMP/rpm-tmp.37236 + cd /Users/doug/tools/RPM/TMP + cd freetype-2.3.5 + rm -rf /Users/doug/tools/RPM/TMP/freetype-2.3.5-root + /Users/doug/tools/bin/make --no-print-directory install DESTDIR=/Users/doug/tools/RPM/TMP/freetype-2.3.5-root make: *** No rule to make target `install'. Stop. error: Bad exit status from /Users/doug/tools/RPM/TMP/rpm-tmp.37236(%install) RPM build errors: Bad exit status from /Users/doug/tools/RPM/TMP/rpm-tmp.37236 (%install)
openpkg bootstrap requires autoconf?
I noticed on recent openpkg bootstrap packages I receive build failures due to missing autoconf. I am building on various versions of solaris and I receive the same error on all. The exact error message is pasted below. ... config.status: creating pathnames.h config.status: creating config.h config.status: executing default-1 commands config.status: creating po/POTFILES config.status: creating po/Makefile config.status: executing default commands + /tmp/openpkg-20070914/make-3.81/make cd . autoconf /bin/sh: autoconf: not found make: *** [configure] Error 1 + exit 2 + exit 2 ./openpkg.boot:ERROR: script returned non-null value
Re: openpkg bootstrap requires autoconf?
Nope, I have all time sync'd machines. I have probably tried it on 8 different sun machines (various OS versions 7,8, and 10). I get this error on every one of them. To get around it I just build m4 and autoconf to get me through the bootstrap. Something else that is interesting, I don't get this error if I update openpkg, it only occurs when performing a bootstrap. On 9/27/07, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Thu, Sep 27, 2007, Ralf S. Engelschall wrote: On Wed, Sep 26, 2007, Doug Henry wrote: I noticed on recent openpkg bootstrap packages I receive build failures due to missing autoconf. I am building on various versions of solaris and I receive the same error on all. The exact error message is pasted below. ... config.status: creating pathnames.h config.status: creating config.h config.status: executing default-1 commands config.status: creating po/POTFILES config.status: creating po/Makefile config.status : executing default commands + /tmp/openpkg-20070914/make-3.81/make cd . autoconf /bin/sh: autoconf: not found make: *** [configure] Error 1 + exit 2 + exit 2 ./openpkg.boot:ERROR: script returned non-null value Strange. That's GNU bash here. Unfortunately I cannot repeat this myself and there are even a few touch(1) calls in openpkg.spec to prevent this. I've to digg deeper... Just a wild guess: can it be that the systems you are trying under have a system time set which is far away from the current real time? Perhaps no ntpd(8) running and so make(1) is confused by some wrong timestamps, etc? Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
gcc conflicts
for distro compatibility I usually test against many versions of gcc, so I figured I would install them from openpkg since they are all available. It seems on 64-bit (I don't remember having this problem on 32-bit) there are naming conflicts, preventing the installation of multiple gcc packages. Processing files: gcc41-4.1.2-20070517 Wrote: /tools/gutsy-64/RPM/PKG/gcc41-4.1.2-20070517.amd64-ubuntu7.10-bsi.rpm Executing(%clean): env -i /tools/gutsy-64/lib/openpkg/bash --norc --noprofile --posix -e /tools/gutsy-64/RPM/TMP/rpm-tmp.73640 + cd /tools/gutsy-64/RPM/TMP + cd gcc-4.1.2 + rm -rf /tools/gutsy-64/RPM/TMP/gcc41-4.1.2-root + exit 0 Executing(--clean): env -i /tools/gutsy-64/lib/openpkg/bash --norc --noprofile --posix -e /tools/gutsy-64/RPM/TMP/rpm-tmp.73640 + cd /tools/gutsy-64/RPM/TMP + rm -rf gcc-4.1.2 + exit 0 Preparing...### [100%] file /tools/gutsy-64/lib64/libstdc++.a from install of gcc41-4.1.2-20070517 conflicts with file from package gcc3-3.4.6-20061013 file /tools/gutsy-64/lib64/libstdc++.la from install of gcc41-4.1.2-20070517 conflicts with file from package gcc3-3.4.6-20061013 file /tools/gutsy-64/lib64/libsupc++.a from install of gcc41-4.1.2-20070517 conflicts with file from package gcc3-3.4.6-20061013 file /tools/gutsy-64/lib64/libsupc++.la from install of gcc41-4.1.2-20070517 conflicts with file from package gcc3-3.4.6-20061013
Re: sqlite fails using with_readline
I found the problem. The readline package is not correctly defining the environment to find the openpkg ncurses install. The following changes to readline.spec must be made to build readline properly: CPPFLAGS=%{l_cppflags ncurses} \ LDFLAGS=%{l_ldflags} \ On 10/30/07, Doug Henry [EMAIL PROTECTED] wrote: I don't have readline installed on my system, I installed the openpkg readline as enforced by sqlite package. I tracked it down to the readline package, I receive the following output to pkg-config --libs readline -L/tools/gutsy-64/lib -lreadline -ltermcap I added --with-curses to the readline spec file and I now receive: -L/tools/gutsy-64/lib -lreadline -lcurses This is not correct either, as libcurses does not exist on my system. I'm working to figure out why the readline package does not find the ncurses library. On 10/30/07, Ralf S. Engelschall [EMAIL PROTECTED] wrote: On Tue, Oct 30, 2007, Doug Henry wrote: Using the with_readline option in sqlite causes a build failure. The termcap library is used (looks like an inclusion by sqlite not openpkg). creating libsqlite3.la (cd .libs rm -f libsqlite3.la ln -s ../libsqlite3.la libsqlite3.la) ./libtool --mode=link /data/los8/tools/dapper-64/bin/cc -fPIC -DNDEBUG=1 -DSQLITE_MAX_SQL_LENGTH=1000 -I. -I./src -DNDEBUG -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -L/data/los8/ tools/dapper-64/lib -DHAVE_READLINE=1 -I/data/los8/tools/dapper-64/include -lpthread \ -o sqlite3 ./src/shell.c libsqlite3.la \ -L/data/los8/tools/dapper-64/lib -lreadline -ltermcap /data/los8/tools/dapper-64/bin/cc -fPIC -DNDEBUG=1 -DSQLITE_MAX_SQL_LENGTH= 1000 -I. -I./src -DNDEBUG -DSQLITE_THREADSAFE=1 -DSQLITE_THREAD_OVERRIDE_LOCK=-1 -DSQLITE_OMIT_LOAD_EXTENSION=1 -DHAVE_READLINE =1 -I/data/los8/tools/dapper-64/include -o sqlite3 ./src/shell.c -L/data/los8/ tools/dapper-64/lib ./.libs/libsqlite3.a -lpthread -lreadline -ltermcap /data/los8/tools/dapper-64/bin/ld: cannot find -ltermcap collect2: ld returned 1 exit status make: *** [sqlite3] Error 1 The sqlite package uses what the readline.pc file from the readline package tells it to link against. There -ltermcap seems to be determined for your platform. Please check there to locate the problem on your platform. Ralf S. Engelschall [EMAIL PROTECTED] www.engelschall.com __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
rsync access
I have been rsync-ing the openpkg current repo for offline use ( rsync.openpkg.org/openpkg-ftp/current/SRC/). Looks like rsync access went away for the paths changed, I get the following error now when trying to rsync: rsync: failed to connect to rsync.openpkg.org: Connection refused (111) Is there new information I can get to continue mirroring? Thanks.
Re: rsync access
Script works nicely...thanks! On Sat, Jan 16, 2010 at 11:41 AM, Armin B. Resch res...@sbcglobal.netwrote: Yes, my mirror broke, too. A post from Thomas in December confirmed rsync and ftp will be turned off. In truth, though, plain rsync without more sophisticated options won’t gain you much in through-put over HTTP because the names of the SRPM files are versioned. Fwiw, here’s the script that I scheduled yesterday to continue mirroring: -- #!/bin/bash SRCDIR=absolute_path_to_your_repo_containing_the_index_file LOGDIR=absolute_path_to_a_log_directory CURL=absolute_path_to_your_curl_executable ts=`date +%Y%m%d` log=$LOGDIR/sync_mirror-$ts.log SITE=http://download.openpkg.org TREE=packages/current/source IDX=00INDEX.rdf.bz2 cd $SRCDIR # Update index echo == Updating index=== \ $log 21 $CURL $SITE/$TREE/$IDX $IDX # Update repo for node in BASE CORE EVAL PLUS;do echo === Processing node $node === \ $log 21 cd $node for FILE in \ `$CURL $SITE/$TREE/$node/ 21 | grep 'class=file' | \ sed 's/.*td valign//' | sed 's/img .*//' | \ sed 's/.*href//' | sed 's/.*\(.*\).*/\1/'`;do PKG=`echo $FILE | sed 's/-[0-9].*//'` oldpkg=`ls $PKG-[0-9]* 2/dev/null` if [ $oldpkg = ];then echo NEW: $PKG $log 21 $CURL $SITE/$TREE/$node/$FILE $FILE 2$log else if [ $oldpkg != $FILE ];then echo Updating package $PKG: old: $oldpkg new: $FILE \ $log 21 rm $oldpkg 2$log $CURL $SITE/$TREE/$node/$FILE $FILE 2$log else echo No change to package $PKG $log 21 fi fi done cd .. done - -- *From:* openpkg-users-ow...@openpkg.org [mailto: openpkg-users-ow...@openpkg.org] *On Behalf Of *Doug Henry *Sent:* Friday, January 15, 2010 9:36 AM *To:* openpkg-users@openpkg.org *Subject:* rsync access I have been rsync-ing the openpkg current repo for offline use ( rsync.openpkg.org/openpkg-ftp/current/SRC/). Looks like rsync access went away for the paths changed, I get the following error now when trying to rsync: rsync: failed to connect to rsync.openpkg.org: Connection refused (111) Is there new information I can get to continue mirroring? Thanks.
openpkg 4 question about rpmmacros
It appears that the new openpkg 4 does not honor the variables set in ~/.rpmmacros file anymore. Is there a new way to set these variables? thanks.
Re: openpkg 4 question about rpmmacros
I didn't set anything special before, it must have searched your home folder by default. I will have a look at your suggested file and see what I find. Thanks. On Wed, Jan 20, 2010 at 12:41 PM, Jeff Johnson n3npq@gmail.com wrote: On Jan 20, 2010, at 9:22 AM, Doug Henry wrote: It appears that the new openpkg 4 does not honor the variables set in ~/.rpmmacros file anymore. Is there a new way to set these variables? You likely have lost whatever envvar's are needed/used for '~' lookup. hth 73 de Jeff__ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
problems compiling libxml correctly on solaris 10
I have an interesting problem on my solaris 10 box I thought someone may be able to help with. The libxml package builds and installs but it does not install all the files, since the build doesn't fail it probably makes it difficult to notice this failure during testing. Doing rpm -ql libxml I get the following listing, which is not even close to a complete libxml install. /tools/sparc10/bin/xml2-config /tools/sparc10/bin/xmlcatalog /tools/sparc10/bin/xmllint /tools/sparc10/include/libxml2 /tools/sparc10/include/libxml2/libxml /tools/sparc10/include/libxml2/libxml/SAX.h /tools/sparc10/include/libxml2/libxml/SAX2.h /tools/sparc10/lib/libxml2.a /tools/sparc10/lib/libxml2.la /tools/sparc10/lib/pkgconfig /tools/sparc10/lib/pkgconfig/libxml-2.0.pc /tools/sparc10/lib/xml2Conf.sh /tools/sparc10/man/man1/xml2-config.1 /tools/sparc10/man/man1/xmllint.1 /tools/sparc10/man/man3/libxml.3 /tools/sparc10/share/aclocal /tools/sparc10/share/aclocal/libxml.m4 I did a rpm -bb and checked in the RPM/TMP folder and poked around with the build, if I run make install I notice the following error during install, which I believe somehow causes the other files to not be installed (I don't see this error on my linux box). /bin/bash ../../mkinstalldirs /tools/sparc10/share/doc/libxml2-2.7.6/html ../.././install-sh -c -m 0644 ./*.html ./*.c ./*.xml ./*.xsl ./*.res /tools/sparc10/share/doc/libxml2-2.7.6/html install: ./*.html does not exist make[3]: [install-data-local] Error 1 (ignored) I noticed the configure/makefile system has a /usr/bin/genhtml program listed, but that executable does not exist on the solaris 10 box, probably has something to do with it. I anyone has any info on what might be causing this problem, please shot me a note. Thanks. -doug
openpkg 4 update and dependencies
I noticed recently when doing openpkg -Ua|sh packages are updating, but the dependencies are not rebuilding, leaving many packages broken. I noticed this most recently with the openssl-1.0.0 release, nothing on my system which depends on openssl rebuilt, openssl simply updated. Running the uninstall command for openssl gives me a list of packages needing openssl, I expect many of which should have been rebuilt with the update (subversion and poco should have been at a minimum). openpkg rpm -e openssl error: Failed dependencies: openssl is needed by (installed) poco-1.3.2-20081023 openssl is needed by (installed) wget-1.12-20091019 openssl is needed by (installed) subversion-1.6.9-20100129 openssl = 0.9.7 is needed by (installed) curl-7.20.0-20100322 openssl is needed by (installed) git-1.7.0.3-20100322 I am mirroring the openpkg directories and adding my own stuff, then using openpkg index to build new INDEX files, which I have done for quite some time. I can't say if this has always been a problem with openpkg 4 or just the most recent 4.0.3 update as I have not been running openpkg 4 for very long. As an example, when jpeg changed to v8 from v6, nothing depending on jpeg rebuilt, causing massive failure when linking libraries built against the old v6 stuff which were not rebuilt during the v8 update. If anyone has an idea of what would cause openpkg to ignore rebuilds caused by dependencies during an update, I would love to try and work this out. Thanks.
Re: Bootstraping on x86/solaris 10 fails
I have always had good luck using the following commands (I just bootstrapped 4.0.5 to make sure nothing funny has happened lately): bash export PATH=/usr/sfw/bin:/usr/ccs/bin:$PATH sh openpkg-4.0.x-x.src.sh --prefix=/opkg without knowing what your paths are its hard to tell if that is the problem, but with the above set I have never had a problem. Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax) On Mon, Mar 8, 2010 at 1:49 PM, Daniel Vergien d...@vergien.net wrote: Hi, I tried to install openpkg like in the tutorial, but it fails: make[3]: Entering directory /tmp/openpkg-4.0.2/rpm-5.1.9/tools' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory /tmp/openpkg-4.0.2/rpm-5.1.9' make: *** [all] Error 2 + exit 2 ./openpkg.boot:ERROR: script returned non-null value -bash-3.00# -bash-3.00# cat /etc/release Solaris 10 5/09 s10x_u7wos_08 X86 Copyright 2009 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 30 March 2009 -bash-3.00# The system is a fresh installed zone. The installscript is openpkg-4.0.2-20100131.src.sh Regards Daniel Vergien __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
Re: openpkg 4 update and dependencies
I'm definitely not using any exclude options and I'm definitely doing -Ua. I have installed a few packages using the --nodeps flag because they pick up strange deps under openpkg 4 (see openpkg-import output below). This is probably worthy of another bug report. Not sure if that gets stored somehow and ends up confusing the update. I didn't use the nodeps option on anything related to openssl, so I wouldn't think it should apply to my case. Executing(--clean): env -i /tools/lin64-testing/lib/openpkg/bash --norc --noprofile --posix -e /tools/lin64-testing/RPM/TMP/rpm-tmp.58500 + cd /tools/lin64-testing/RPM/TMP error: Failed dependencies: /usr/bin/addr2line is needed by openpkg-import-0-20070904 /usr/bin/ar is needed by openpkg-import-0-20070904 /usr/bin/as is needed by openpkg-import-0-20070904 /usr/bin/gprof is needed by openpkg-import-0-20070904 /usr/bin/ld is needed by openpkg-import-0-20070904 /usr/bin/nm is needed by openpkg-import-0-20070904 /usr/bin/objcopy is needed by openpkg-import-0-20070904 /usr/bin/objdump is needed by openpkg-import-0-20070904 /usr/bin/ranlib is needed by openpkg-import-0-20070904 /usr/bin/readelf is needed by openpkg-import-0-20070904 /usr/bin/size is needed by openpkg-import-0-20070904 /usr/bin/strings is needed by openpkg-import-0-20070904 Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax) On Fri, Apr 2, 2010 at 2:43 PM, Ralf S. Engelschall r...@openpkg.org wrote: On Thu, Apr 01, 2010, Doug Henry wrote: I noticed recently when doing openpkg -Ua|sh packages are updating, but the dependencies are not rebuilding, leaving many packages broken. I noticed this most recently with the openssl-1.0.0 release, nothing on my system which depends on openssl rebuilt, openssl simply updated. Running the uninstall command for openssl gives me a list of packages needing openssl, I expect many of which should have been rebuilt with the update (subversion and poco should have been at a minimum). openpkg rpm -e openssl error: Failed dependencies: openssl is needed by (installed) poco-1.3.2-20081023 openssl is needed by (installed) wget-1.12-20091019 openssl is needed by (installed) subversion-1.6.9-20100129 openssl = 0.9.7 is needed by (installed) curl-7.20.0-20100322 openssl is needed by (installed) git-1.7.0.3-20100322 Are you really sure you are using -Ua and not -Uaq? Because your behaviour is what the q option enforces. Without q the depending packages should be rebuild just fine. I am mirroring the openpkg directories and adding my own stuff, then using openpkg index to build new INDEX files, which I have done for quite some time. I can't say if this has always been a problem with openpkg 4 or just the most recent 4.0.3 update as I have not been running openpkg 4 for very long. As an example, when jpeg changed to v8 from v6, nothing depending on jpeg rebuilt, causing massive failure when linking libraries built against the old v6 stuff which were not rebuilt during the v8 update. If anyone has an idea of what would cause openpkg to ignore rebuilds caused by dependencies during an update, I would love to try and work this out. I currently have no clue what happens here. Usually with -Ua everything rebuilds just fine and the behavior you see results only under -Uaq. Hmmm... are there perhaps some -E (exclude) options present? Or this there an .openpkg/build script under ~%{l_musr} which contains -q or something like this? Ralf S. Engelschall r...@engelschall.com www.engelschall.com __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
Re: Bootstraping on x86/solaris 10 fails
I am able to replicate this exact bootstrap problem on my sparc 8 box with openpkg-4.0.5 libtool: link: /usr/local/bin/gcc -D_GNU_SOURCE -D_REENTRANT -o rpm2cpio rpm2cpio.o -L/tmp/openpkg-4.0.5/uuid-1.6.2/.libs -L/tmp/openpkg-4.0.5/pcre-8.01/.libs -L/tmp/openpkg-4.0.5/sqlite-3.6.17/.libs -L/tmp/openpkg-4.0.5/beecrypt-4.2.1/.libs -L/tmp/openpkg-4.0.5/bzip2-1.0.5/.libs -L/tmp/openpkg-4.0.5/popt-1.15/.libs -L/tmp/openpkg-4.0.5/popt-1.15 -L/tmp/openpkg-4.0.5/zlib-1.2.3 -L/tmp/openpkg-4.0.5/bzip2-1.0.5 -L/tmp/openpkg-4.0.5/beecrypt-4.2.1 -L/tmp/openpkg-4.0.5/openssl-0.9.8k/lib -L/tmp/openpkg-4.0.5/sqlite-3.6.17 -L/tmp/openpkg-4.0.5/pcre-8.01 -L/tmp/openpkg-4.0.5/uuid-1.6.2 ../lib/.libs/librpm.a -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/sparc-sun-solaris2.8/libstdc++-v3/src -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/sparc-sun-solaris2.8/libstdc++-v3/src/.libs -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/gcc -L/usr/local/sparc-sun-solaris2.8/bin -L/usr/local/sparc-sun-solaris2.8/lib -L/usr/local/lib/../sparc-sun-solaris2.8/lib -L/usr/local/lib -L/usr/ccs/bin -L/usr/ccs/lib ../rpmdb/.libs/librpmdb.a ../rpmio/.libs/librpmio.a ../misc/.libs/librpmmisc.a -L/tmp/openpkg-4.0.5/rpm-5.1.9/db3 -L/tmp/openpkg-4.0.5/rpm-5.1.9/lua -lresolv /tmp/openpkg-4.0.5/uuid-1.6.2/.libs/libuuid.a /tmp/openpkg-4.0.5/pcre-8.01/.libs/libpcreposix.a /tmp/openpkg-4.0.5/pcre-8.01/.libs/libpcre.a /tmp/openpkg-4.0.5/sqlite-3.6.17/.libs/libsqlite3.a -lcrypto /tmp/openpkg-4.0.5/beecrypt-4.2.1/.libs/libbeecrypt.a -ldl /usr/local/lib/libstdc++.so -lgcc_s /tmp/openpkg-4.0.5/bzip2-1.0.5/.libs/libbz2.a -lz /tmp/openpkg-4.0.5/popt-1.15/.libs/libpopt.a -lrt -lsocket -lnsl -lm -R/usr/local/lib -R/usr/local/lib Undefined first referenced symbol in file libiconv_open ../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv_close ../rpmdb/.libs/librpmdb.a(hdrfmt.o) ld: fatal: Symbol referencing errors. No output written to rpm2cpio collect2: ld returned 1 exit status Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax) On Sat, Apr 3, 2010 at 7:33 AM, Daniel Vergien d...@vergien.net wrote: On 02.04.10 20:40, Ralf S. Engelschall wrote: This is too less information. What errors occur in the tools directory of RPM? I There has to be more details... Oops, copied to less output... libtool: link: /usr/sfw/bin/gcc -D_GNU_SOURCE -D_REENTRANT -o rpm2cpio rpm2cpio. o -L/tmp/openpkg-4.0.2/uuid-1.6.2/.libs -L/tmp/openpkg-4.0.2/pcre-8.01/.libs -L /tmp/openpkg-4.0.2/sqlite-3.6.17/.libs -L/tmp/openpkg-4.0.2/beecrypt-4.2.1/.libs -L/tmp/openpkg-4.0.2/bzip2-1.0.5/.libs -L/tmp/openpkg-4.0.2/popt-1.15/.libs -L/ tmp/openpkg-4.0.2/popt-1.15 -L/tmp/openpkg-4.0.2/zlib-1.2.3 -L/tmp/openpkg-4.0.2 /bzip2-1.0.5 -L/tmp/openpkg-4.0.2/beecrypt-4.2.1 -L/tmp/openpkg-4.0.2/openssl-0. 9.8k/lib -L/tmp/openpkg-4.0.2/sqlite-3.6.17 -L/tmp/openpkg-4.0.2/pcre-8.01 -L/tm p/openpkg-4.0.2/uuid-1.6.2 ../lib/.libs/librpm.a -L/usr/sfw/lib ../rpmdb/.libs/l ibrpmdb.a ../rpmio/.libs/librpmio.a ../misc/.libs/librpmmisc.a -L/tmp/openpkg-4. 0.2/rpm-5.1.9/db3 -L/tmp/openpkg-4.0.2/rpm-5.1.9/lua -lresolv /tmp/openpkg-4.0.2 /uuid-1.6.2/.libs/libuuid.a /tmp/openpkg-4.0.2/pcre-8.01/.libs/libpcreposix.a /t mp/openpkg-4.0.2/pcre-8.01/.libs/libpcre.a -ldl /tmp/openpkg-4.0.2/sqlite-3.6.17 /.libs/libsqlite3.a -lcrypto /tmp/openpkg-4.0.2/beecrypt-4.2.1/.libs/libbeecrypt .a /usr/sfw/lib/libstdc++.so -lgcc_s /tmp/openpkg-4.0.2/bzip2-1.0.5/.libs/libbz2 .a -lz /tmp/openpkg-4.0.2/popt-1.15/.libs/libpopt.a -lrt -lsocket -lnsl -lm -R/u sr/sfw/lib -R/usr/sfw/lib Undefined first referenced symbol in file libiconv_close ../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv_open ../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv../rpmdb/.libs/librpmdb.a(hdrfmt.o) ld: fatal: Symbol referencing errors. No output written to rpm2cpio collect2: ld returned 1 exit status make[3]: *** [rpm2cpio] Error 1 make[3]: Leaving directory `/tmp/openpkg-4.0.2/rpm-5.1.9/tools' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/tmp/openpkg-4.0.2/rpm-5.1.9/tools' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/tmp/openpkg-4.0.2/rpm-5.1.9' make: *** [all] Error 2 + exit 2 ./openpkg.boot:ERROR: script returned non-null value -bash-3.00# echo $PATH /usr/sfw/bin:/usr/ccs/bin:/usr/sbin:/usr/bin __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
more goofy dependency problems
I get a lot of this stuff with openpkg 4, not sure why or what it means... + /tools/lin64-testing/lib/openpkg/rpmtool files -v -ofiles -r/tools/lin64-testing/RPM/TMP/openssl-1.0.0-20100331-buildroot '%defattr(-,tools,tools)' /tools/lin64-testing '%not %dir {/tools/lin64-testing,/tools/lin64-testing/*,/tools/lin64-testing/etc/rc.d,/tools/lin64-testing/man/*}' '%config /tools/lin64-testing/etc/openssl/openssl.cnf' rpmtool:files: pass 1 (preparation and syntactical expansions) rpmtool:files: pass 2 (filesystem-based expansions) rpmtool:files: pass 3 (duplication removal and cleanup) + exit 0 Processing files: openssl-1.0.0-20100331 Wrote: /tools/lin64-testing/RPM/PKG/openssl-1.0.0-20100331.amd64-ubuntu6.06-bsi.rpm Executing(%clean): env -i /tools/lin64-testing/lib/openpkg/bash --norc --noprofile --posix -e /tools/lin64-testing/RPM/TMP/rpm-tmp.19048 + cd /tools/lin64-testing/RPM/TMP + cd openssl-1.0.0 + rm -rf /tools/lin64-testing/RPM/TMP/openssl-1.0.0-20100331-buildroot Executing(--clean): env -i /tools/lin64-testing/lib/openpkg/bash --norc --noprofile --posix -e /tools/lin64-testing/RPM/TMP/rpm-tmp.19048 + cd /tools/lin64-testing/RPM/TMP + rm -rf openssl-1.0.0 error: Failed dependencies: /tools/lin64-testing/lib64 is needed by openssl-1.0.0-20100331 Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax)
Re: openpkg 4 update and dependencies
sorry...the command would be openpkg build -Ua Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax) On Tue, Apr 6, 2010 at 3:42 PM, Jeff Johnson n3npq@gmail.com wrote: On Apr 6, 2010, at 11:33 AM, Doug Henry wrote: I'm definitely not using any exclude options and I'm definitely doing -Ua. I have installed a few packages using the --nodeps flag because they pick up strange deps under openpkg 4 (see openpkg-import output below). This is probably worthy of another bug report. Not sure if that gets stored somehow and ends up confusing the update. I didn't use the nodeps option on anything related to openssl, so I wouldn't think it should apply to my case. Executing(--clean): env -i /tools/lin64-testing/lib/openpkg/bash --norc --noprofile --posix -e /tools/lin64-testing/RPM/TMP/rpm-tmp.58500 + cd /tools/lin64-testing/RPM/TMP error: Failed dependencies: /usr/bin/addr2line is needed by openpkg-import-0-20070904 /usr/bin/ar is needed by openpkg-import-0-20070904 /usr/bin/as is needed by openpkg-import-0-20070904 /usr/bin/gprof is needed by openpkg-import-0-20070904 /usr/bin/ld is needed by openpkg-import-0-20070904 /usr/bin/nm is needed by openpkg-import-0-20070904 /usr/bin/objcopy is needed by openpkg-import-0-20070904 /usr/bin/objdump is needed by openpkg-import-0-20070904 /usr/bin/ranlib is needed by openpkg-import-0-20070904 /usr/bin/readelf is needed by openpkg-import-0-20070904 /usr/bin/size is needed by openpkg-import-0-20070904 /usr/bin/strings is needed by openpkg-import-0-20070904 These are all executable dependencies (although I cannot tell where they are coming from). Installing binutils will supply all those paths. Why binutils is not automagically included is a different question ... (aside) Hmmm -Ua is no meaningful RPM option (although likely is accepted w/o any error message). What did you intend with -Ua? 73 de Jeff __ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org
Re: Bootstraping on x86/solaris 10 fails
I think the solution to this problem...at least on my sparc 8 box...was to: export LD_LIBRARY_PATH=/usr/local/lib (or some path where the libiconv stuff was installed...on my box the sunfreeware gcc and libiconv installed to /usr/local) Doug Henry Brilligent Solutions Inc. 937.545.0609 (cell) 937.879.4158 (fax) On Tue, Apr 6, 2010 at 3:34 PM, Jeff Johnson n3npq@gmail.com wrote: On Apr 6, 2010, at 11:48 AM, Doug Henry wrote: I am able to replicate this exact bootstrap problem on my sparc 8 box with openpkg-4.0.5 libtool: link: /usr/local/bin/gcc -D_GNU_SOURCE -D_REENTRANT -o rpm2cpio rpm2cpio.o -L/tmp/openpkg-4.0.5/uuid-1.6.2/.libs -L/tmp/openpkg-4.0.5/pcre-8.01/.libs -L/tmp/openpkg-4.0.5/sqlite-3.6.17/.libs -L/tmp/openpkg-4.0.5/beecrypt-4.2.1/.libs -L/tmp/openpkg-4.0.5/bzip2-1.0.5/.libs -L/tmp/openpkg-4.0.5/popt-1.15/.libs -L/tmp/openpkg-4.0.5/popt-1.15 -L/tmp/openpkg-4.0.5/zlib-1.2.3 -L/tmp/openpkg-4.0.5/bzip2-1.0.5 -L/tmp/openpkg-4.0.5/beecrypt-4.2.1 -L/tmp/openpkg-4.0.5/openssl-0.9.8k/lib -L/tmp/openpkg-4.0.5/sqlite-3.6.17 -L/tmp/openpkg-4.0.5/pcre-8.01 -L/tmp/openpkg-4.0.5/uuid-1.6.2 ../lib/.libs/librpm.a -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/sparc-sun-solaris2.8/libstdc++-v3/src -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/sparc-sun-solaris2.8/libstdc++-v3/src/.libs -L/sol10/SOURCES/S8/gcc-3.4.6/objdir/gcc -L/usr/local/sparc-sun-solaris2.8/bin -L/usr/local/sparc-sun-solaris2.8/lib -L/usr/local/lib/../sparc-sun-solaris2.8/lib -L/usr/local/lib -L/usr/ccs/bin -L/usr/ccs/lib ../rpmdb/.libs/librpmdb.a ../rpmio/.libs/librpmio.a ../misc/.libs/librpmmisc.a -L/tmp/openpkg-4.0.5/rpm-5.1.9/db3 -L/tmp/openpkg-4.0.5/rpm-5.1.9/lua -lresolv /tmp/openpkg-4.0.5/uuid-1.6.2/.libs/libuuid.a /tmp/openpkg-4.0.5/pcre-8.01/.libs/libpcreposix.a /tmp/openpkg-4.0.5/pcre-8.01/.libs/libpcre.a /tmp/openpkg-4.0.5/sqlite-3.6.17/.libs/libsqlite3.a -lcrypto /tmp/openpkg-4.0.5/beecrypt-4.2.1/.libs/libbeecrypt.a -ldl /usr/local/lib/libstdc++.so -lgcc_s /tmp/openpkg-4.0.5/bzip2-1.0.5/.libs/libbz2.a -lz /tmp/openpkg-4.0.5/popt-1.15/.libs/libpopt.a -lrt -lsocket -lnsl -lm -R/usr/local/lib -R/usr/local/lib Undefined first referenced symbol in file libiconv_open ../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv../rpmdb/.libs/librpmdb.a(hdrfmt.o) libiconv_close ../rpmdb/.libs/librpmdb.a(hdrfmt.o) ld: fatal: Symbol referencing errors. No output written to rpm2cpio collect2: ld returned 1 exit status The actual usage case for -liconv is rather teensy and easily corrected if desired. In order to automate i18n/l10n conversions as part of --query, rpm has something called header format extensions. E.g., the is a local conversion tied to this --queryformat rpm -qa --queryformat '%{name:iconv}' If you you don't _WANT_ that automated iconv functionality, its easily stubbed out or #ifdef'd depending on whether -liconv is available (or not). hth 73 de Jeff__ OpenPKG http://openpkg.org User Communication List openpkg-users@openpkg.org