cd ../../unxlngppc.pro/lib && regcomp -register -br uno_types.rdb -r uno_services.rdb\ -c acceptor.uno.so \ -c bridgefac.uno.so \ -c connector.uno.so \ -c remotebridge.uno.so \ -c uuresolver.uno.so \ -c bridgetest.uno.so \ -c cppobj.uno.so \ -c uriproc.uno.so register component 'acceptor.uno.so' in registry 'uno_services.rdb' succesful! register component 'bridgefac.uno.so' in registry 'uno_services.rdb' succesful! register component 'connector.uno.so' in registry 'uno_services.rdb' succesful! register component 'remotebridge.uno.so' in registry 'uno_services.rdb' succesful! register component 'uuresolver.uno.so' in registry 'uno_services.rdb' succesful!register component 'bridgetest.uno.so' in registry 'uno_services.rdb' succesful!register component 'cppobj.uno.so' in registry 'uno_services.rdb' succesful! register component 'uriproc.uno.so' in registry 'uno_services.rdb' succesful! regcomp -register -br ../../unxlngppc.pro/lib/uno_types.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb \ -c javaloader.uno.so -c javavm.uno.so register component 'javaloader.uno.so' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! register component 'javavm.uno.so' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c \ file:///var/tmp/portage/openoffice-2.0.1-r1/work/ooo-build-2.0.1.3/build/ooa680-m1/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar \ -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.1-r1/work/ooo-build-2.0.1.3/build/ooa680-m1/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 [Java framework]sunjavaplugin.so could not load Java runtime library: file:///opt/ibm-jdk-bin-1.4.2.05/jre/bin/classic/libjvm.so.register component 'file:///var/tmp/portage/openoffice-2.0.1-r1/work/ooo-build-2.0.1.3/build/ooa680-m1/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' failed! error (CannotRegisterImplementationException): Could not create Java implementation loader dmake: Error code 1, while making '../../unxlngppc.pro/lib/uno_services.rdb' dmake: '../../unxlngppc.pro/lib/uno_services.rdb' removed. '---* tg_merge.mk *---' ERROR: Error 65280 occurred while making /var/tmp/portage/openoffice-2.0.1-r1/work/ooo-build-2.0.1.3/build/ooa680-m1/testtools/source/bridgetest make: *** [stamp/build] Error 1 !!! ERROR: app-office/openoffice-2.0.1-r1 failed. Call stack: ebuild.sh, line 1531: Called dyn_compile ebuild.sh, line 931: Called src_compile openoffice-2.0.1-r1.ebuild, line 214: Called die !!! Build failed !!! If you need support, post the topmost build error, and the call stack if relevant. My emerge --info: emerge --info Portage 2.1_rc3-r3 (default-linux/ppc/ppc32/2006.0/G4, gcc-4.1.1, glibc-2.4-r3, 2.6.16-ck11 ppc) ================================================================= System uname: 2.6.16-ck11 ppc 7455, altivec supported Gentoo Base System version 1.12.0 dev-lang/python: 2.2.3-r6, 2.3.5-r2, 2.4.3-r1 dev-python/pycrypto: 2.0.1-r5 dev-util/ccache: [Not Present] dev-util/confcache: [Not Present] sys-apps/sandbox: 1.2.18.1 sys-devel/autoconf: 2.13, 2.59-r7 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2 sys-devel/binutils: 2.16.1-r2 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.11-r4 ACCEPT_KEYWORDS="ppc ~ppc" AUTOCLEAN="yes" CBUILD="powerpc-unknown-linux-gnu" CFLAGS="-O2 -mcpu=7450 -maltivec -mabi=altivec -pipe -ftree-vectorize -ftree-vectorizer-verbose=1 -fPIC -fno-strict-aliasing" CHOST="powerpc-unknown-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/lib/mozilla/defaults/pref /usr/share/X11/xkb /usr/share/config /var/bind" CONFIG_PROTECT_MASK="/etc/gconf /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c /etc/env.d" CXXFLAGS="-O2 -mcpu=7450 -maltivec -mabi=altivec -pipe -ftree-vectorize -ftree-vectorizer-verbose=1 -fPIC -fno-strict-aliasing" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig buildpkg distlocks metadata-transfer sandbox sfperms strict" GENTOO_MIRRORS="http://gentoo.osuosl.org/ http://mirror.usu.edu/mirrors/gentoo/ http://mirror.mcs.anl.gov/pub/gentoo/ http://gentoo.inode.at/ http://trumpetti.atm.tut.fi/gentoo/ http://linux.rz.ruhr-uni-bochum.de/download/gentoo-mirror/ http://ftp.heanet.ie/pub/gentoo/ http://ftp.easynet.nl/mirror/gentoo/ http://ftp.snt.utwente.nl/pub/os/linux/gentoo http://linuv.uv.es/mirror/gentoo/ http://mirror.switch.ch/ftp/mirror/gentoo/ http://ibiblio.org/pub/Linux/MIRRORS.html " LANG="es_ES.UTF-8" LC_ALL="es_ES.UTF-8" LINGUAS="es_ES es en_US en_GB" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude='/distfiles' --exclude='/local' --exclude='/packages'" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage /usr/local/xgl-coffee /usr/local/xgl-testing" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="ppc X aac accessibility alsa altivec apache2 apm arts avahi berkdb bitmap-fonts bluetooth bonobo bzip2 cairo cdr cjk cli crypt cups curl dbus dri dv dvd dvdr dvdread eds emacs emboss encode esd fbcon firefox foomaticdb fortran gdbm gif glitz gmp gnome gpm gstreamer gtk gtk2 gtkhtml hal icq imlib insecure-savers ipv6 isdnlog jabber java jpeg jython kde lcms ldap libg++ libwww mad mikmod mono motif mozcalendar mozilla mozsvg mp3 mpeg ncurses nls nptl ogg opengl pam pcmcia pcre pdflib perl pic png pnp pppd python qt quicktime readline reflection ruby sasl sdl session spell spl ssl svg tcpd theora threads truetype truetype-fonts type1-fonts udev unicode usb v4l v4l2 vorbis xinerama xml xml2 xmms xorg xosd xv xvid yahoo zeroconf zlib dvb_cards_usb-wt220u dvb_cards_usb-dtt200u elibc_glibc input_devices_keyboard input_devices_mouse kernel_linux linguas_es_ES linguas_es linguas_en_US linguas_en_GB userland_GNU video_cards_ati video_cards_radeon" Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LDFLAGS, PORTAGE_RSYNC_EXTRA_OPTS
I forgot to mention that I needed to add a # define _STLP_STRICT_ANSI 1 in /var/tmp/portage/openoffice-2.0.1-r1/work/ooo-build-2.0.1.3/build/OOO_2_0_1/stlport/unxlngppc.pro/misc/build/STLport-4.5/stlport/config/stl_gcc.h or the respective file in 2.0.2* to avoid a compilation error with gcc-4.1.1 Not sure if anybody else sees this problem, or if it depends on CFLAGS.
Reopen if you can reproduce the thing after you've removed stuff like "-ftree-vectorize -ftree-vectorizer-verbose=1 -fPIC -fno-strict-aliasing" from your C[XX]FLAGS.
first, -O2 (forced by the ebuild) implies -ftree-vectorize second, -ftree-vectorizer-verbose is just dumping where it did something, not causing any effect -fPIC should be harmless, in fact I had to use it because buggy ebuilds would generate .so without it, which is bad and gives problems in ppc ditto for -fno-strict-aliasing, in fact it is in make.conf.example in portage, and was mandatory for ppc for a long time.
(In reply to comment #3) > first, -O2 (forced by the ebuild) implies -ftree-vectorize No, it doesn't, check man gcc. In fact, it's pretty broken still, search bugzilla for related bugs. > -fPIC should be harmless, in fact I had to use it because buggy ebuilds would > generate .so without it, which is bad and gives problems in ppc No, that's definitely not harmless and you shouldn't stick such stuff into make.conf at all. It's just hiding bugs and breaking things that expects -dPIC as well. Look, reproduce the problem w/ sane C[XX]FLAGS and reopen then.
you're right WRT -ftree-vectorize, I was misled by the way -ftree-vect-loop-version and -ftree-vrp is written in the gcc-4.1.1 manual. I did some experiments, and empirical evidence is on your side :) The compile is proceeding further, I sidestepped both errors, I'll test with the CFLAGS you mean if it does not build with the current ones. WRT the "STRICT_ANSI" stuff, it should be independent of CXXFLAGS, it is a language syntax stuff, that gets triggered in stlport by gcc-4.1.1
I hope this time you accept it: cd ./unxlngppc.pro/misc/build/STLport-4.5/src && make -f gcc-3.0.mak -j1 && touch so_built_so_stlport make[1]: Entering directory `/var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/stlport/unxlngppc.pro/misc/build/STLport-4.5/src' mkdir -p ../lib/obj/GCCppc/ReleaseD g++ -Wl,-Bdirect -D_REENTRANT -DGXX_INCLUDE_PATH=/usr/lib/gcc/powerpc-unknown-linux-gnu/4.1.1/include/g++-v4 -fexceptions -I../stlport -Wall -W -Wno-sign-compare -Wno-unused -Wno-uninitialized -ftemplate-depth-32 -O2 -O2 -fPIC dll_main.cpp -c -o ../lib/obj/GCCppc/ReleaseD/dll_main.o ../stlport/cstdio:111: error: '__std_alias::vsnprintf' has not been declared make[1]: *** [../lib/obj/GCCppc/ReleaseD/dll_main.o] Error 1 make[1]: Leaving directory `/var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/stlport/unxlngppc.pro/misc/build/STLport-4.5/src' dmake: Error code 2, while making './unxlngppc.pro/misc/build/so_built_so_stlport' '---* tg_merge.mk *---' ERROR: Error 65280 occurred while making /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/stlport make: *** [stamp/build] Error 1 !!! ERROR: app-office/openoffice-2.0.2-r2 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile openoffice-2.0.2-r2.ebuild, line 235: Called die !!! Build failed !!! If you need support, post the topmost build error, and the call stack if relevant. with emerge --info: sgala@marlow ~ $ emerge --info Portage 2.1_rc4-r1 (default-linux/ppc/ppc32/2006.0/G4, gcc-4.1.1, glibc-2.4-r3, 2.6.16-ck11 ppc) ================================================================= System uname: 2.6.16-ck11 ppc 7455, altivec supported Gentoo Base System version 1.12.0 dev-lang/python: 2.2.3-r6, 2.3.5-r2, 2.4.3-r1 dev-python/pycrypto: 2.0.1-r5 dev-util/ccache: [Not Present] dev-util/confcache: [Not Present] sys-apps/sandbox: 1.2.18.1 sys-devel/autoconf: 2.13, 2.59-r7 sys-devel/automake: 1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2 sys-devel/binutils: 2.16.1-r2 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.11-r4 ACCEPT_KEYWORDS="ppc ~ppc" AUTOCLEAN="yes" CBUILD="powerpc-unknown-linux-gnu" CFLAGS="-O2" CHOST="powerpc-unknown-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/3.4/env /usr/kde/3.4/share/config /usr/kde/3.4/shutdown /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/lib/mozilla/defaults/pref /usr/share/X11/xkb /usr/share/config /var/bind" CONFIG_PROTECT_MASK="/etc/gconf /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c /etc/env.d" CXXFLAGS="-O2" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig buildpkg distlocks metadata-transfer sandbox sfperms strict" GENTOO_MIRRORS="http://gentoo.osuosl.org/ http://mirror.usu.edu/mirrors/gentoo/ http://mirror.mcs.anl.gov/pub/gentoo/ http://gentoo.inode.at/ http://trumpetti.atm.tut.fi/gentoo/ http://linux.rz.ruhr-uni-bochum.de/download/gentoo-mirror/ http://ftp.heanet.ie/pub/gentoo/ http://ftp.easynet.nl/mirror/gentoo/ http://ftp.snt.utwente.nl/pub/os/linux/gentoo http://linuv.uv.es/mirror/gentoo/ http://mirror.switch.ch/ftp/mirror/gentoo/ http://ibiblio.org/pub/Linux/MIRRORS.html " LANG="es_ES.UTF-8" LC_ALL="es_ES.UTF-8" LINGUAS="es_ES es en_US en_GB" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --delete-after --stats --timeout=180 --exclude='/distfiles' --exclude='/local' --exclude='/packages'" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage /usr/local/xgl-coffee /usr/local/xgl-testing" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="ppc X aac accessibility alsa altivec apache2 apm arts avahi berkdb bitmap-fonts bluetooth bonobo browserplugin bzip2 cairo cdr cjk cli crypt cups curl dbus dri dv dvd dvdr dvdread eds emacs emboss encode esd fbcon firefox foomaticdb fortran gdbm gif glitz gmp gnome gpm gstreamer gtk gtk2 gtkhtml hal icq imlib insecure-savers ipv6 isdnlog jabber java jpeg jython kde lcms ldap libg++ libwww mad mikmod mono motif mozcalendar mozilla mozsvg mp3 mpeg ncurses nls nptl ogg opengl pam pcmcia pcre pdflib perl pic png pnp pppd python qt quicktime readline reflection ruby sasl sdl session spell spl ssl svg tcpd theora threads truetype truetype-fonts type1-fonts udev unicode usb v4l v4l2 vorbis xinerama xml xml2 xmms xorg xosd xv xvid yahoo zeroconf zlib dvb_cards_usb-wt220u dvb_cards_usb-dtt200u elibc_glibc input_devices_keyboard input_devices_mouse kernel_linux linguas_es_ES linguas_es linguas_en_US linguas_en_GB userland_GNU video_cards_ati video_cards_radeon" Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LDFLAGS, PORTAGE_RSYNC_EXTRA_OPTS The workaround is easy for this one: --- /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/stlport/unxlngppc.pro/misc/build/STLport-4.5/stlport/config/stl_gcc.h 2006-06-04 13:38:00.000000000 +0200 +++ /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/stlport/unxlngppc.pro/misc/build/STLport-4.5/stlport/config/stl_gcc.h.mio 2006-06-04 13:37:19.000000000 +0200 @@ -226,7 +226,7 @@ # endif # endif /* _STLP_WEAK_ATTRIBUTE */ -# define _STLP_STRICT_ANSI 1 + /* strict ANSI prohibits "long long" ( gcc) */ # if defined ( __STRICT_ANSI__ ) # undef _STLP_LONG_LONG which should be probably conditionalized on >=gcc-4 (0 or 1, not sure). done, restarted ebuild, waiting for the next failure
re: comment #2 probably removing the line: append-flags "-fno-strict-aliasing" from openoffice-2.0.1-r1 (current stable ebuild) would make sense, as you are ignoring bug requests including it.
(In reply to comment #7) > re: comment #2 > > probably removing the line: > append-flags "-fno-strict-aliasing" > > from openoffice-2.0.1-r1 (current stable ebuild) would make sense, as you are > ignoring bug requests including it. > Nobody is ignoring anything, what are you talking about? Anyway, that's a case for the ppc-herd, as it works fine on x86 and can't reproduce this as I don't have the necessary hardware. Still in advance, the first of your bugs seems like a regcomp-problem, which is indeed normally caused by CFLAGS-overoptimization.
The regcomp thing, now without *any* CFLAGS except -O2 (as seen in my emerge --info: cd ../../unxlngppc.pro/lib && regcomp -register -br uno_types.rdb -r uno_services.rdb\ -c acceptor.uno.so \ -c bridgefac.uno.so \ -c connector.uno.so \ -c remotebridge.uno.so \ -c uuresolver.uno.so \ -c bridgetest.uno.so \ -c cppobj.uno.so \ -c uriproc.uno.so \ -c constructors.uno.so register component 'acceptor.uno.so' in registry 'uno_services.rdb' succesful! register component 'bridgefac.uno.so' in registry 'uno_services.rdb' succesful! register component 'connector.uno.so' in registry 'uno_services.rdb' succesful! register component 'remotebridge.uno.so' in registry 'uno_services.rdb' succesful! register component 'uuresolver.uno.so' in registry 'uno_services.rdb' succesful!register component 'bridgetest.uno.so' in registry 'uno_services.rdb' succesful!register component 'cppobj.uno.so' in registry 'uno_services.rdb' succesful! register component 'uriproc.uno.so' in registry 'uno_services.rdb' succesful! register component 'constructors.uno.so' in registry 'uno_services.rdb' succesful! regcomp -register -br ../../unxlngppc.pro/lib/uno_types.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb \ -c javaloader.uno.so -c javavm.uno.so register component 'javaloader.uno.so' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! register component 'javavm.uno.so' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c \ file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar \ -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 register component 'file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' failed! error (RuntimeException): JNI exception occurred dmake: Error code 1, while making '../../unxlngppc.pro/lib/uno_services.rdb' dmake: '../../unxlngppc.pro/lib/uno_services.rdb' removed. '---* tg_merge.mk *---' ERROR: Error 65280 occurred while making /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest make: *** [stamp/build] Error 1 !!! ERROR: app-office/openoffice-2.0.2-r2 failed. Call stack: ebuild.sh, line 1539: Called dyn_compile ebuild.sh, line 939: Called src_compile openoffice-2.0.2-r2.ebuild, line 235: Called die !!! Build failed !!! If you need support, post the topmost build error, and the call stack if relevant. it took some time arriving here from the 4.1.1 error.
I tried using different java VMs in the step that fails, and Sun's evaluation JRE was a bit more helpful than the other: regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c \ file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar \ -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 Unexpected Signal sig 11, dumping state ... Registers: 0 [0x0ffbe30c] [0x7f9e2668] [0x3024df10] [0x0eb82a18] 4 [0x3884ffd5] [0x00000000] [0x0fd6df10] [0x00000018] 8 [0x00000000] [0x0eb99134] [0x00000030] [0x3001f0a0] 12 [0x48044482] [0x37a04108] [0x7f9e2750] [0x1002c600] 16 [0x7f9e2734] [0x7f9e2738] [0x7f9e2964] [0x10131fa0] 20 [0x00000001] [0x00000000] [0x7f9e2773] [0x0000000c] 24 [0x7f9e2730] [0x300039f0] [0x0eb9a320] [0x7f9e2694] 28 [0x7f9e2698] [0x0eb82a18] [0x0ffc5c60] [0x30003a1c] Frame: sp 0x7f9e2668 fp 0x7f9e2678 pc 0xfe0aaf4 0x7f9e2668 [0x7f9e2678] [0x300039f0] [0x0eb9a320] [0x7f9e2694] Instructions: 0x0fe0aad4 [0x4bffffc0] 0x0fe0aad8 [0x7fe3fb78] 0x0fe0aadc [0x3be00000] 0x0fe0aae0 [0x4bffd90d] 0x0fe0aae4 [0x4bffffbc] 0x0fe0aae8 [0x9421fff0] 0x0fe0aaec [0x7c801828] 0x0fe0aaf0 [0x38840001] 0x0fe0aaf4 [0x7c80192d] 0x0fe0aaf8 [0x40a2fff4] 0x0fe0aafc [0x4c00012c] 0x0fe0ab00 [0x7c832378] 0x0fe0ab04 [0x38210010] 0x0fe0ab08 [0x4e800020] 0x0fe0ab0c [0x9421fff0] 0x0fe0ab10 [0x7c801828] # # An unexpected error has been detected by HotSpot Virtual Machine: # # SIGSEGV (0xb) at pc=0x0fe0aaf4, pid=11824, tid=807692976 # # Java VM: Java HotSpot(TM) Client VM (1.4.2_11-eval-b12 mixed mode) # Problematic frame: # C [libuno_sal.so.3+0x1baf4] osl_incrementInterlockedCount+0xc # # An error report file with more information is saved as hs_err_pid11824.log # # If you would like to submit a bug report, please visit: # http://java.sun.com/webapps/bugreport/crash.jsp # Abort question is: this functions is written in assembly in /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/sal/osl/unx/interlck.c can anyone with experience in C/assembly take a look in case there is something obvious there?
To get a more verbose JNI exception (with Java trace) for ibm-jdk use you could apply: http://go-oo.org/patches/src680/jni-throw-exception.diff --- The SIGSEGV happens if libgcc3_uno.so is built with '-fomit-frame-pointer'. At least for x86. This is also not reproducible. On some machines it happens on others not. That with identical arch/cpu! Well, just to note. I know '-fomit-frame-pointer' flag is filtered. May be gcc-4.1.1 enables it with -'O2' now (for ppc)?! To verify that it does or not you could apply: http://go-oo.org/patches/src680/omit-fp-workaround.diff It adds '-mno-omit-leaf-frame-pointer -fno-omit-frame-pointer' to build libgcc3_uno.so. So if '-O2' implies it these cut it off for sure.
I tried with your patches, and I'm getting something interesting: exactly the same result with ibm-jdk-bin-1.4.2.05: regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c \ file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar \ -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 [Java framework]sunjavaplugin.so could not load Java runtime library: file:///opt/ibm-jdk-bin-1.4.2.05/jre/bin/classic/libjvm.so.register component 'file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' failed! error (CannotRegisterImplementationException): Could not create Java implementation loader dmake: Error code 1, while making '../../unxlngppc.pro/lib/uno_services.rdb' dmake: '../../unxlngppc.pro/lib/uno_services.rdb' removed. '---* tg_merge.mk *---' No backtrace, same failure. Now, when (after this failure), I do: sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ echo $JAVA_HOME /opt/ibm-jdk-bin-1.4.2.05 sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ export JAVA_HOME=/home/sgala/j2re1.4.2_11-eval sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 register component 'file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ dmake dmake: Executing shell macro: -+ls -1 $(JARDIR) | $(GREP) "^$i" dmake: Executing shell macro: -+ls -1 $(JARDIR) | $(GREP) "^$i" dmake: Executing shell macro: -+ls -1 $(JARDIR) | $(GREP) "^$i" dmake: Executing shell macro: $(FIND) . -name "{$(subst,$($(WINVERSIONNAMES)_MAJOR),* $(subst,$(UPD)$(DLLPOSTFIX), $(SHL1TARGET)))}.xml" dmake: Executing shell macro: $(FIND) . -name "{$(subst,$($(WINVERSIONNAMES)_MAJOR),* $(subst,$(UPD)$(DLLPOSTFIX), $(SHL2TARGET)))}.xml" dmake: Executing shell macro: $(FIND) . -name "{$(subst,$($(WINVERSIONNAMES)_MAJOR),* $(subst,$(UPD)$(DLLPOSTFIX), $(SHL3TARGET)))}.xml" ------------- cp -p bridgetest_inprocess ../../unxlngppc.pro/lib/bridgetest_inprocess chmod +x ../../unxlngppc.pro/lib/bridgetest_inprocess cp -p bridgetest_server ../../unxlngppc.pro/lib/bridgetest_server chmod +x ../../unxlngppc.pro/lib/bridgetest_server cp -p bridgetest_client ../../unxlngppc.pro/lib/bridgetest_client chmod +x ../../unxlngppc.pro/lib/bridgetest_client rm -f ../../unxlngppc.pro/lib/bridgetest_javaserver echo java -classpath /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin/java_uno.jar:/var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin/ridl.jar:/var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin/jurt.jar:/var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin/juh.jar:/opt/ibm-jdk-bin-1.4.2.05/jre/lib/rt.jar:.:../class:../class/testComponent.jar \ com.sun.star.comp.bridge.TestComponentMain \ \""uno:socket,host=127.0.0.1,port=2002;urp;test"\" \ > ../../unxlngppc.pro/lib/bridgetest_javaserver chmod +x ../../unxlngppc.pro/lib/bridgetest_javaserver rm -f ../../unxlngppc.pro/lib/bridgetest_inprocess_java echo uno -ro uno_services.rdb -ro uno_types.rdb \ -s com.sun.star.test.bridge.BridgeTest \ -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin \ -- com.sun.star.test.bridge.JavaTestObject > ../../unxlngppc.pro/lib/bridgetest_inprocess_java chmod +x ../../unxlngppc.pro/lib/bridgetest_inprocess_java i.e. switching to the Sun "evaluation" embedded JDK succeeds. Before, it gave the error about Hotspot that I posted. All is very strange.
And this is with ibm-jdk-bin-1.5.0 (SR2). It uses the j9VM: export JAVA_HOME=/opt/ibm-jdk-bin-1.5.0 sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 Exception in thread "main" java.lang.ClassNotFoundException: com.sun.star.lib.unoloader.UnoClassLoader at java.net.URLClassLoader$ClassFinder.run(URLClassLoader.java:963) at java.security.AccessController.doPrivileged(AccessController.java:274) at java.net.URLClassLoader.findClass(URLClassLoader.java:487) at java.lang.ClassLoader.loadClass(ClassLoader.java:604) at java.lang.ClassLoader.loadClass(ClassLoader.java:561) Caused by: java.util.zip.ZipException: invalid code lengths set at java.util.zip.InflaterInputStream.read(InflaterInputStream.java:165) at sun.misc.Resource.getBytes(Resource.java:115) at java.net.URLClassLoader.defineClass(URLClassLoader.java:552) at java.net.URLClassLoader.access$400(URLClassLoader.java:119) at java.net.URLClassLoader$ClassFinder.run(URLClassLoader.java:961) ... 4 more register component 'file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' failed! error (RuntimeException): JNI exception occurred sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ export JAVA_HOME=/home/sgala/j2re1.4.2_11-eval sgala@marlow /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest $ regcomp -register -br ../../unxlngppc.pro/misc/bridgetest/bootstrap.rdb -r ../../unxlngppc.pro/lib/uno_services.rdb -c file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar -env:URE_INTERNAL_JAVA_DIR=file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/solver/680/unxlngppc.pro/bin using loader com.sun.star.loader.Java2 register component 'file:///var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/testtools/source/bridgetest/../../unxlngppc.pro/class/testComponent.jar' in registry '../../unxlngppc.pro/lib/uno_services.rdb' succesful! huh? (I can't understand it, really)
(In reply to comment #13) > And this is with ibm-jdk-bin-1.5.0 (SR2). It uses the j9VM: > ibm-jdk-bin-1.5.0 is known to break: http://bugs.gentoo.org/show_bug.cgi?id=126105
@Andreas That is fixed in SR1 of ibm-jdk-1.5.x @Santiago To summarise: 1. ibm-jdk-1.4.x error This is due to bad CFLAGS?! You talked with Jakub. Another thing you could check if 'strace solver/680/unx....pro/bin/javaldx' succeeds or fails. And if not why. 2. sun-jdk success gcc-4.1.1 enables -fomit-frame-pointer which the second patch workaround. 3. ibm-jdk-1.5.x error Caused by: java.util.zip.ZipException: invalid code lengths set Now how to find out what the CLASSPATH looks like. ibm-jdk-1.5.x could just skip broken items in CLASSPATH. Well, I just guess that this is a CLASSPATH issue. But I saw such issues with sun-jdk fixed a long time ago.
@Andreas See this bugreport at eclipse.org: https://bugs.eclipse.org/bugs/show_bug.cgi?id=115269
(In reply to comment #15) > @Andreas > That is fixed in SR1 of ibm-jdk-1.5.x > Which unfortunately is not in the tree...
I can reproduce comment #6 on PPC. Adding a -D_STLP_STRICT_ANSI in the CFLAGS does seem to fix it and gets me through the openoffice-2.0.2-r2 compile on PowerPC (With GCC4.1)
I am using ibm-jdk-bin 1.5.0 SR2 (local ebuild) and 1.4.5.05 (again local) I noticed that the patch got overwritten, the one you gave me for -fno-omit-frame-pointer. (BTW, -fno-omit-leaf-frame-pointer is not legal in powerpc) I applied the patch again, rebuilt libgcc3_uno.so, copied it to solver..., and the results are the same. I'm still building after regcomp succeeded with the Sun jdk for embedded whatever. I'll report when it finishes (I need to CTRL-Z it from time to time, it is hot currently in Spain)
I'm trying to summarize after a private exchange of emails with Hanno Meyer-Thurow and getting a mostly working OO.o with java: - -D_STLP_STRICT_ANSI (either in CXXFLAGS or patching somewhere) enables stl, and thus full compilation using gcc-4.1.1 (this covers the title of this bug report, the rest seems to be about UNO/java related problems). I guess the maintainer should either rename this bug after patching the issue, or close this bug and let me open another one) - gcc-4.1.1 enables -fomit-frame-pointer with -O2, at least. SIGSEGV and other UNO problems happen if libgcc3_uno.so is built with '-fomit-frame-pointer'. http://go-oo.org/patches/src680/omit-fp-workaround.diff tries to avoid it, and seems needed - still, I had problems with IBM jdk-1.4.2, which I circumvented using source /var/tmp/portage/openoffice-2.0.2-r2/work/ooo-build-2.0.2.11/build/OOO_2_0_2/LinuxPPCEnv.Set* (or modifying parts of it to use the Sun jre for embedded to run regcomp) bot in testtools and during the final install. My working copy is so dirty that I'd say: - consolidate the _STL_STRICT_ANSI stuff either in CXXFLAGS or patched - consolidate the -fno-omit-frame-pointer either in CXXFLAGS or patched Try again the new build, and check further problems with ibm-jdk-1.4.2.*
(In reply to comment #20) > - -D_STLP_STRICT_ANSI (either in CXXFLAGS or patching somewhere) enables stl, > and thus full compilation using gcc-4.1.1 So i guess something like use ppc && append-flags "-D_STLP_STRICT_ANSI" should be sufficient, right? > - gcc-4.1.1 enables -fomit-frame-pointer with -O2, at least. SIGSEGV and other > UNO problems happen if libgcc3_uno.so is built with '-fomit-frame-pointer'. > http://go-oo.org/patches/src680/omit-fp-workaround.diff tries to avoid it, and > seems needed Actually we filter -fomit-frame-pointer in the 2.0.2-ebuilds, so I'm not sure why this happens with -r2.
(In reply to comment #21) > Actually we filter -fomit-frame-pointer in the 2.0.2-ebuilds, so I'm not sure > why this happens with -r2. Again, the optimization level '-O2' implies it. There is no need to add '-fomit-frame-pointer' to CFLAGS then. ;)
(In reply to comment #22) > (In reply to comment #21) > > Actually we filter -fomit-frame-pointer in the 2.0.2-ebuilds, so I'm not sure > > why this happens with -r2. > > Again, the optimization level '-O2' implies it. > There is no need to add '-fomit-frame-pointer' to CFLAGS then. ;) > Thanks, I can read ;) but that's not what I said... We FILTER not ADD it in the ebuild atm. It's just quite odd, that it's obviously not filtered just because it's implied on a plattform. Anyway: I've added both fixes to the -r2 ebuild now, first through adding to the CFLAGS and second by using Hannos patch.
Comment #21 >So i guess something like > >use ppc && append-flags "-D_STLP_STRICT_ANSI" > >should be sufficient, right? I guess it is a gcc-4.1 issue, but works for me. Cool, thanks. I hope they will advance stlport version upstream, or patch it. It would be really great if we could reuse icu, stlport, jpeg, db, ... from the distribution ebuilds, but I guess it is really a problem with a distribution oriented to source builds. I think I'll reopen the java related problems when I have a clearer picture of it. The title (building with 4.1) is reported to work for a couple users.