| Summary: | Openoffice 2.0.4 fails to install on ppc: unable to register pyUNO modules | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Luca Barbato <lu_zero> |
| Component: | New packages | Assignee: | Gentoo Office Team <office> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | dberkholz, ed, frp.bissey, gent_bz, joho, michael, neil, ppc, pylon |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | PPC | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 159951 | ||
|
Description
Luca Barbato
2006-09-14 02:30:08 UTC
Portage 2.1.1 (default-linux/ppc/ppc32/2006.0/G4, gcc-4.1.1, glibc-2.4-r3, 2.6.17-gentoo-r4 ppc) ================================================================= System uname: 2.6.17-gentoo-r4 ppc 7455, altivec supported Gentoo Base System version 1.12.1 Last Sync: Wed, 13 Sep 2006 16:50:01 +0000 distcc 2.18.3 powerpc-unknown-linux-gnu (protocols 1 and 2) (default port 3632) [disabled] app-admin/eselect-compiler: [Not Present] dev-lang/python: 2.3.5, 2.4.2-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-r1 sys-devel/binutils: 2.17 sys-devel/gcc-config: 1.3.13-r1 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=G4 -mtune=G4 -fno-strict-aliasing -pipe" CHOST="powerpc-unknown-linux-gnu" As much as I would like it to: No, pyuno is needed. Still this is usually triggered by some specific CFLAGS, so maybe if we find out what is safe, we could filter this. This seems to be a continuation of bug #139412, please especially see comments 26 and 27. And just to say it loud: I really do hate pyuno / regcomp Just changing the summary... Mine: CFLAGS="-O2 -march=G4 -maltivec -mabi=altivec -fno-strict-aliasing -pipe -g -w" (In reply to comment #4) > Mine: > CFLAGS="-O2 -march=G4 -maltivec -mabi=altivec -fno-strict-aliasing -pipe -g -w" > In the bug I mentioned above someone states, that -march=G4 or G3 won't work, while -march=7400 does. Also on another bug someone mentioned that replacing -O2 with just -O helps. So it would be great if you could try this out, note that for the second you would have to add an replace-flag line to the ebuild, as we atm are filtering all optimization levels to just -O2 as this until now is the only one working reliably (at least on other archs...) Some input from someone who has more knowledge about the ppc-arch (if there would be a better solution) would also be great. (In reply to comment #5) > Also on another bug someone mentioned that replacing > -O2 with just -O helps. So it would be great if you could try this out, note > that for the second you would have to add an replace-flag line to the ebuild, > as we atm are filtering all optimization levels to just -O2 as this until now > is the only one working reliably (at least on other archs...) Actually I looked at this myself and you do not filter -O, just -O?. Anyhow I'll kick off a 7400 build and see whether that helps. Should know in ~20 hours. *** Bug 148458 has been marked as a duplicate of this bug. *** (In reply to comment #6) > > Actually I looked at this myself and you do not filter -O, just -O?. Anyhow > I'll kick off a 7400 build and see whether that helps. Should know in ~20 > hours. > Any news on this? I built a working ooo dropping G4 and using 7400 hadn't time to investigate. I was not aware that create_services_rdb is that pyuno-bug, that's because I created bug #148458. Anyways, I can build OOo with the CFLAGS from the ppc32-profile, that means, with CFLAGS="-O2 -mcpu=powerpc -mtune=powerpc -fno-strict-aliasing -pipe" But I want to mention, that I built the complete environment with that CFLAGS. Currently I try to build and install OOo with -mcpu=750 -mtune=750 instead of G3. But in an environment that has been built with -mcpu=G3 -mtune=G3 previously. If that fails I'll rebuild the complete environment with the 750-settings. Probably there is a dependency (python?) which fails with the G3 or G4 optimization? (In reply to comment #10) > I was not aware that create_services_rdb is that pyuno-bug, that's because I > created bug #148458. Anyways, I can build OOo with the CFLAGS from the > ppc32-profile, that means, with CFLAGS="-O2 -mcpu=powerpc -mtune=powerpc > -fno-strict-aliasing -pipe" > > But I want to mention, that I built the complete environment with that CFLAGS. > Currently I try to build and install OOo with -mcpu=750 -mtune=750 instead of > G3. But in an environment that has been built with -mcpu=G3 -mtune=G3 > previously. If that fails I'll rebuild the complete environment with the > 750-settings. Probably there is a dependency (python?) which fails with the G3 > or G4 optimization? > If that fails it indeed most likely is a problem with python, we had quite a similar problem when python is compiled with --fast-math recently (In reply to comment #10) > But I want to mention, that I built the complete environment with that CFLAGS. > Currently I try to build and install OOo with -mcpu=750 -mtune=750 instead of > G3. But in an environment that has been built with -mcpu=G3 -mtune=G3 > previously. Okay, that fails. Now to the second try with rebuilding the complete environment with the changed CFLAGS. (In reply to comment #11) > If that fails it indeed most likely is a problem with python, we had quite a > similar problem when python is compiled with --fast-math recently We have a ppc-meeting this Sunday. And one of the topics is a discussion if we need to change our default CFLAGS from G3/G4 to 750/7400, caused by this and probably some other bugs (JoseJX can tell more). (In reply to comment #12) > (In reply to comment #10) > > But I want to mention, that I built the complete environment with that CFLAGS. > > Currently I try to build and install OOo with -mcpu=750 -mtune=750 instead of > > G3. But in an environment that has been built with -mcpu=G3 -mtune=G3 > > previously. > > Okay, that fails. Now to the second try with rebuilding the complete > environment with the changed CFLAGS. > You really first might try to only recompile python with the lowered CFLAGS so that we can nail this down to where it exactly orginates from... (In reply to comment #14) > You really first might try to only recompile python with the lowered CFLAGS so > that we can nail this down to where it exactly orginates from... No luck... Rebuilding stages and the GRP-environment now. (In reply to comment #8) > (In reply to comment #6) > > > > Actually I looked at this myself and you do not filter -O, just -O?. Anyhow > > I'll kick off a 7400 build and see whether that helps. Should know in ~20 > > hours. > > > > Any news on this? No sorry, shortly after that my hard drive got corrupted, I had to reinstall, and I haven't had time to play on it. I did confirm, however, that dropping the CFLAGS on just python does nothing. So I suspect it's OOO itself that's fragile, or perhaps the combination of G4/O2 CFLAGS on both OOO and python. (In reply to comment #16) > (In reply to comment #8) > > (In reply to comment #6) > > > > > > Actually I looked at this myself and you do not filter -O, just -O?. Anyhow > > > I'll kick off a 7400 build and see whether that helps. Should know in ~20 > > > hours. > > > > > > > Any news on this? > > No sorry, shortly after that my hard drive got corrupted, I had to reinstall, > and I haven't had time to play on it. I did confirm, however, that dropping the > CFLAGS on just python does nothing. So I suspect it's OOO itself that's > fragile, or perhaps the combination of G4/O2 CFLAGS on both OOO and python. > Yeah, seems I wasn't to clear on that, I meant reducing the CFLAGS for BOTH python and OOo, only python won't be enough (at least that's what other cases showed) I rebuilt the complete stages and GRP-environment with -mcpu=750 -mtune=750 -- without luck. OOo still fails with ERROR: ERROR: Could not register all components! in function: create_services_rdb Adding ppc@ to Cc. Probably JoseJX can comment, as he said, that switching from G4 to 7400 for mcpu and mtune helped for him. (In reply to comment #18) > I rebuilt the complete stages and GRP-environment with -mcpu=750 -mtune=750 -- > without luck. OOo still fails with > ERROR: ERROR: Could not register all components! > in function: create_services_rdb > > Adding ppc@ to Cc. Probably JoseJX can comment, as he said, that switching > from G4 to 7400 for mcpu and mtune helped for him. Try switching -mcpu=7450 to -mcpu=7400. That worked for me, too. Might be solved with rc2. (In reply to comment #20) > Might be solved with rc2. Might... Did you read ChangeLogs? I can give it a try in a day or two. (In reply to comment #20) > Might be solved with rc2. No, it is not solved by RC2, I tried. I have this problem while trying to build openoffice-2.0.3. CFLAGS="-O2 -pipe" (In reply to comment #22) > (In reply to comment #20) > > Might be solved with rc2. > > No, it is not solved by RC2, I tried. Confirmed in my build-environment. It still fails on optimised CFLAGS, but works with ppc-generic CFLAGS. Any progress here? OOo 2.0.4 should be release in a few days and it would be good to have some sort of solution for this... beside turning -mcpu/tune to specific one and not the generic G4 I cannot think of other workarounds... (In reply to comment #26) > beside turning -mcpu/tune to specific one and not the generic G4 I cannot think > of other workarounds... That does not matter. The complete toolchain needs to be built with -mcpu=ppc -tune=ppc. Substituting G3 or G4 with ppc in the CFLAGS for the OpenOffice-ebuild only does not help. Has this issue been reported upstream? I can't see that we can resolve it on our own. using CFLAGS="-O2 -mcpu=7400 -mtune=7400 -fno-strict-aliasing -pipe" just for openoffice let me have a stable build (In reply to comment #28) > using > > CFLAGS="-O2 -mcpu=7400 -mtune=7400 -fno-strict-aliasing -pipe" But not with -mcpu=750 -mtune=750 or -mcpu=G3 -mtune=G3 (In reply to comment #28) > CFLAGS="-O2 -mcpu=7400 -mtune=7400 -fno-strict-aliasing -pipe" And not with CFLAGS="-O2 -mtune=G4 -mcpu=G4 -maltivec -mabi=altivec -pipe" and USE="branding cairo cups dbus eds firefox gnome gstreamer gtk kde ldap mono pam sound webdav -binfilter -debug -java -odk" LINGUAS="de en -af -ar -be_BY -bg -bn -bs -ca -cs -cy -da -el -en_GB -en_US -en_ZA -es -et -fa -fi -fr -gu_IN -he -hi_IN -hr -hu -it -ja -km -ko -lt -mk -nb -nl -nn -nr -ns -pa_IN -pl -pt -pt_BR -ru -rw -sh_YU -sk -sl -sr_CS -st -sv -sw_TZ -th -tn -tr -ts -vi -xh -zh_CN -zh_TW -zu" I just searched for the error again and it seems that Ubuntu found a solution: https://launchpad.net/distros/ubuntu/+source/openoffice.org/+bug/62590 I'll test it. (In reply to comment #31) > I just searched for the error again and it seems that Ubuntu found a solution: > https://launchpad.net/distros/ubuntu/+source/openoffice.org/+bug/62590 > > I'll test it. > Looks like a different problem to me, the log there is complaining about pythonloader and not pyuno Ok, for the short term (as OOo 2.0.4 should be release in the next days), I' m going to add some replace-action to the ebuild like: # problems with regcomp segfaulting during install, see bug #147542 if use ppc; then replace-flags "-march=g4" "-march=7400" replace-flags "-march=g3" "-march=7400" replace-flags "-march=750" "-march=7400" replace-flags "-mtune=g4" "-mtune=7400" replace-flags "-mtune=g3" "-mtune=7400" replace-flags "-mtune=750" "-mtune=7400" fi But for the long term I agree with pylon - this has to be solved upstream. So could someone who has the necessary hardware to test file a bug upstream (http://qa.openoffice.org/issues) and post the bug number here, so I can add myself to the CC: ? g3 line is wrong... (In reply to comment #34) > g3 line is wrong... > How so? (In reply to comment #32) > Looks like a different problem to me, the log there is complaining about > pythonloader and not pyuno You're right. And it does not fix our problem. (In reply to comment #33) > Ok, for the short term (as OOo 2.0.4 should be release in the next days), I' m > going to add some replace-action to the ebuild like: I doubt that this will work. I just started an installation with -mcpu=7400 -mtune=7400. Let's see tomorrow if that helps... > # problems with regcomp segfaulting during install, see bug #147542 > if use ppc; then For all lines below: We don't use -march on ppc. We have -mcpu and -mtune only. > replace-flags "-march=g3" "-march=7400" > replace-flags "-march=750" "-march=7400" > replace-flags "-mtune=g4" "-mtune=7400" > replace-flags "-mtune=g3" "-mtune=7400" > replace-flags "-mtune=750" "-mtune=7400" > replace-flags "-march=g4" "-march=7400" ppc = ppc g3 = 750 g4 = 7400 There are some more aliases around, but I guess, that quite everybody will use g3 or g4, as this is the default for us. That is what replace-cpu-flags() is for ... also note G3, G4, not g3, g4. (In reply to comment #28) > using > > CFLAGS="-O2 -mcpu=7400 -mtune=7400 -fno-strict-aliasing -pipe" > > just for openoffice let me have a stable build Fails for me (although I don't have -fno-strict-aliasing set): ------------------------------ Making: ../../unxlngppc.pro/slo/cf_service.obj g++ -fsigned-char -fmessage-length=0 -c -O2 -fno-strict-aliasing -D_FILE_OFFSET_BITS=64 -I. -I../../unxlngppc.pro/inc/canvasfactory.uno -I../inc -I../../inc/pch -I../../inc -I../../unx/inc -I../../unxlngppc.pro/inc -I. -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solver/680/unxlngppc.pro/inc/stl -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solver/680/unxlngppc.pro/inc/external -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solver/680/unxlngppc.pro/inc -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solenv/unxlngppc/inc -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solenv/inc -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/res -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solver/680/unxlngppc.pro/inc/stl -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solenv/inc/Xp31 -INO_JAVA_HOME/include -INO_JAVA_HOME/include/linux -INO_JAVA_HOME/include/native_threads/include -I/usr/include -I/var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/solver/680/unxlngppc.pro/inc/offuh -I. -I../../res -I. -fsigned-char -pipe -frtti -O2 -mtune=7400 -mcpu=7400 -maltivec -mabi=altivec -pipe -D_STLP_STRICT_ANSI -fexceptions -fno-enforce-eh-specs -Wno-ctor-dtor-privacy -fPIC -DLINUX -DUNX -DVCL -DGCC -DC300 -DPOWERPC -DCVER=C300 -DNPTL -DGLIBC=2 -D_PTHREADS -D_REENTRANT -DNEW_SOLAR -D_USE_NAMESPACE=1 -DSTLPORT_VERSION=400 -DPOWERPC -DPPC -DHAVE_GCC_VISIBILITY_FEATURE -D__DMAKE -DUNIX -DCPPU_ENV=gcc3 -DGXX_INCLUDE_PATH=/usr/lib/gcc/powerpc-unknown-linux-gnu/4.1.1/include/g++-v4 -DSUPD=680 -DPRODUCT -DNDEBUG -DPRODUCT_FULL -DOSL_DEBUG_LEVEL=0 -DOPTIMIZE -DEXCEPTIONS_ON -DGSTREAMER -DCUI -DOOD680=OOD680 -DSHAREDLIB -D_DLL_ -DMULTITHREAD -o ../../unxlngppc.pro/slo/cf_service.o /var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/canvas/source/factory/cf_service.cxx /var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/canvas/source/factory/cf_service.cxx: In constructor '<unnamed>::CanvasFactory::CanvasFactory(const com::sun::star::uno::Reference<com::sun::star::uno::XComponentContext>&)': /var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/canvas/source/factory/cf_service.cxx:127: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://bugs.gentoo.org/> for instructions. The bug is not reproducible, so it is likely a hardware or OS problem. dmake: Error code 1, while making '../../unxlngppc.pro/slo/cf_service.obj' '---* tg_merge.mk *---' ERROR: Error 65280 occurred while making /var/tmp/portage/openoffice-2.0.4_rc2/work/ooo/build/ood680-m4/canvas/source/factory make: *** [stamp/build] Error 1 (In reply to comment #39) > > Fails for me (although I don't have -fno-strict-aliasing set): That's a totally different error. Is this reproducable? And are you sure that's not some sort of hardware problem? (like overheating or such) Compiler sefaulting mostly is caused by hardware issues. (In reply to comment #38) > That is what replace-cpu-flags() is for ... also note G3, G4, not g3, g4. > Nice, never seen that before ;) so would adding something like replace-cpu-flags g3 g4 750 7400 work or will this mess up stuff for g3-users? In know there were some comments, that the root of the problem lies deeper, but the CFLAG-reduction seems to work for most, so we might use it. But in the end it's the call of the ppc-herd, until then I've removed the ppc-keyword from the openoffice-2.0.4-ebuild (so that it can be unmasked for the other archs) I just like to say that I had no problem (on ppc) with rc1 or rc2. However since 2.0.2 or 2.0.3 I exclusively compile OOo with CFLAGS="" and CXXFLAGS=CFLAGS. No other thing I tried work and OOo configuration provides some optimisation flags all by itself anyway. I achieve this by having a file in "/etc/portage/env/" (namely app-office/openoffice) with the special CFLAGS defined. My emerge info is: Portage 2.1.1-r1 (default-linux/ppc/ppc32/2006.1/G4, gcc-4.1.1, glibc-2.4-r3, 2.6.16-gentoo-r9 ppc) ================================================================= System uname: 2.6.16-gentoo-r9 ppc 7450, altivec supported Gentoo Base System version 1.12.5 Last Sync: Thu, 12 Oct 2006 18:50:01 +0000 ccache version 2.3 [enabled] app-admin/eselect-compiler: [Not Present] dev-java/java-config: 1.2.11-r1 dev-lang/python: 2.4.3-r4 dev-python/pycrypto: 2.0.1-r5 dev-util/ccache: 2.3 dev-util/confcache: [Not Present] sys-apps/sandbox: 1.2.17 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-r3 sys-devel/gcc-config: 1.3.13-r4 sys-devel/libtool: 1.5.22 virtual/os-headers: 2.6.16 ACCEPT_KEYWORDS="ppc" AUTOCLEAN="yes" CBUILD="powerpc-unknown-linux-gnu" CFLAGS="-O2 -mcpu=7450 -mtune=7450 -pipe -maltivec -mabi=altivec -frename-registers -fivopts -ftree-vectorize -fno-strict-aliasing" CHOST="powerpc-unknown-linux-gnu" CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/share/X11/xkb /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/revdep-rebuild /etc/terminfo" CXXFLAGS="-O2 -mcpu=7450 -mtune=7450 -pipe -maltivec -mabi=altivec -frename-registers -fivopts -ftree-vectorize -fno-strict-aliasing" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig ccache distlocks metadata-transfer parallel-fetch sandbox sfperms strict" GENTOO_MIRRORS="ftp://ftp.vic.keypoint.com.au http://mirrors.tds.net/gentoo ftp://mirrors.tds.net/gentoo http://mirror.tucdemonic.org/gentoo/" LC_ALL="en_GB.UTF-8" LINGUAS="fr en_GB" 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/portage/local/layman/science /usr/portage/local/layman/ebuild-exchange" SYNC="rsync://rsync.au.gentoo.org/gentoo-portage" USE="ppc X a52 aac aalib alsa altivec amr arts asf audiofile berkdb bitmap-fonts bonobo bzip2 cairo cddb cdf cdparanoia cdr cli crypt cups curl divx dlloader dri dts dv dvd dvdr dvdread elibc_glibc emboss encode esd expat f77 ffmepg fftw firefox flac foomaticdb fortran gdbm gif gmedia gmp gpm graphviz gs gstreamer gtk gtk2 gtkhtml hal idn ieee1394 imagemagick imlib imlib2 input_devices_evdev input_devices_keyboard input_devices_mouse ipv6 isdnlog java jbig jpeg jpeg2k kde kdehiddenvisibility kdexdeltas kernel_linux lcms libg++ libwww linguas_en_GB linguas_fr live lzo mad mikmod mjpeg mng mp3 mp4 mpeg musepack musicbrainz ncurses netcdf network nls nptl nptlonly ogg openal openexr opengl oss pam pcre pdf perl png ppds pppd python qt3 qt4 quicktime rdesktop readline reflection ruby samba sdl session slang slp sndfile spell spl ssl svg szip tcl tcltk tcpd tetex theora tiff tk truetype truetype-fonts type1-fonts udev unicode usb userland_GNU vcd video_cards_fbdev video_cards_nv vorbis wmf wmp x264 xine xml xml2 xmms xorg xv xvid zeroconf zlib" Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LDFLAGS, MAKEOPTS, PORTAGE_RSYNC_EXTRA_OPTS (In reply to comment #40) > (In reply to comment #39) > > > > Fails for me (although I don't have -fno-strict-aliasing set): > > That's a totally different error. Is this reproducable? And are you sure that's > not some sort of hardware problem? (like overheating or such) Compiler > sefaulting mostly is caused by hardware issues. Sorry for the delay. I thought I already answered to it. It is reproducable when I use -mcpu=7400 -mtune=7400. At least it failed two times with internal compiler error: Segmentation fault I doubt that this Pegasos G4 is overheating or has hardware issues. It just compiled KDE and it can compile OpenOffice until the installation-phase where it hits the PyUNO-bug. Furthermore I tried to compile CFLAGS="-O2 -pipe", where it failed in the install phase once again. *sigh* Seems there is no way to compile OpenOffice with our current system-layout. I can try to rebuild my toolchain with lu_zero's CFLAGS and then compile OpenOffice. I've been doing many tests with oo-2.0.3 (custom ebuild with no CFLAG filtering) and all kinds of CFLAGS. I'm not finished, but I suspect my results will mostly apply to openoffice-2.0.4_rc1 as well. Empirically: Failure: CFLAGS="-O2 -mcpu=750 -mtune=750 -pipe" Success: CFLAGS="" Failure: CFLAGS="-O2 -pipe" Success: CFLAGS="-O1 -pipe -mcpu=750 -mtune=750" Success: CFLAGS="-Os -pipe -mcpu=750 -mtune=750" Failure: CFLAGS="-O2 -pipe -mcpu=750 -mtune=750 -fno-align-functions -fno-align-jumps -fno-align-loops -fno-align-labels" Failure: CFLAGS="-O2 -pipe -mcpu=750 -mtune=750 -fno-reorder-blocks -fno-reorder-blocks-and-partition -fno-prefetch-loop-arrays -fno-tree-vect-loop-version" Those last three are fairly interesting to me. Note that the CFLAGS I turned off constitute the bulk of the difference between -Os (which works) and -O2 (which fails). Yet -O2 with either half of its size-increasing CFLAGS turned off still fails. It's probably useful to know that it builds with -Os, however, so I posted this interim report. *** Bug 153143 has been marked as a duplicate of this bug. *** CFLAGS="-O2 -mcpu=7400 -mtune=7400 -maltivec -fno-strict-aliasing -pipe" worked for me (on a mac mini). CFLAGS="-O2 -mcpu=7400 -mtune=7400 -maltivec -fno-strict-aliasing -pipe" worked for me with (ooo-2.0.3 on a mac mini). I've added a new patch to the ebuild from upstream, which tries to solve the regcomp segfaults. So could you please try out the new version and see if the fix helps in your case? Just finished here. Failure with the flags I tried:
make_installer.pl, version 1.0
Product list file: openoffice.lst
Taking setup script from solver
Unpackpath: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util
Compiler: unxlngppc
Product: OpenOffice
BuildID: 9073
Build: OOD680
No minor set
Product version
Installpath: /usr/lib/openoffice
Package format: native
Package list file: ../inc_openoffice/unix/packagelist.txt
Addon-Package list file: ../inc_openoffice/unix/packagelist_language.txt
Not calling epm
No file stripping
Unzip ARCHIVE files
services.rdb can be created
Languages:
fr
en-GB
########################################################
... checking required files ...
...... searching zip ...
Found: /usr/bin/zip
...... searching unzip ...
Found: /usr/bin/unzip
... analyzing openoffice.lst ...
... analyzing script: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/setup_osl.ins ...
... analyzing directories ...
... analyzing files ...
... analyzing scpactions ...
... analyzing shortcuts ...
... analyzing profile ...
... analyzing profileitems ...
... analyzing modules ...
------------------------------------
... languages fr ...
... analyzing files ...
... analyzing files with flag ARCHIVE ...
... analyzing files with flag SCPZIP_REPLACE ...
... analyzing files with flag PATCH_SO_NAME ...
... creating preregistered services.rdb ...
**************************************************
ERROR: ERROR: Could not register all components!
in function: create_services_rdb
**************************************************
**************************************************
ERROR: Saved logfile: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//logging/fr_en-GB/log_OOD680__fr_en-GB.log
---------------------------------------------------
the relevant part of the log file is the following:
Sun Nov 5 12:30:42 2006 (01:13 min.)
######################################################
Registering python UNO components:
######################################################
SUCCESS: Source for types.rdb: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb
SUCCESS: Source for pyuno_services.rdb: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/p$
Systemcall: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/t$
ERROR: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/po$
Systemcall: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/t$
ERROR: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/po$
Moved directory from /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr$
Removing directory /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//zip/fr_en-GB
Removing directory /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_e$
***************************************************************
ERROR: Could not register all components!
in function: create_services_rdb
***************************************************************
-----------------------------------------------------
and here is my emerge --info
Portage 2.1.1-r1 (default-linux/ppc/ppc32/2006.1/G4, gcc-4.1.1, glibc-2.4-r3, 2.6.16-gentoo-r9 ppc)
=================================================================
System uname: 2.6.16-gentoo-r9 ppc 7450, altivec supported
Gentoo Base System version 1.12.6
Last Sync: Fri, 03 Nov 2006 23:50:01 +0000
ccache version 2.3 [enabled]
app-admin/eselect-compiler: [Not Present]
dev-java/java-config: 1.3.7, 2.0.30
dev-lang/python: 2.4.3-r4
dev-python/pycrypto: 2.0.1-r5
dev-util/ccache: 2.3
dev-util/confcache: [Not Present]
sys-apps/sandbox: 1.2.17
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: 2.17
sys-devel/gcc-config: 1.3.13-r4
sys-devel/libtool: 1.5.22
virtual/os-headers: 2.6.16
ACCEPT_KEYWORDS="ppc"
AUTOCLEAN="yes"
CBUILD="powerpc-unknown-linux-gnu"
CFLAGS="-O2 -mcpu=7450 -mtune=7450 -pipe -maltivec -mabi=altivec -fno-strict-aliasing"
CHOST="powerpc-unknown-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config /usr/kde/3.5/shutdown /usr/share/X11/xkb /usr/share/config /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/xdvi/"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/gconf /etc/java-config/vms/ /etc/revdep-rebuild /etc/terminfo"
CXXFLAGS="-O2 -mcpu=7450 -mtune=7450 -pipe -maltivec -mabi=altivec -fno-strict-aliasing"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache distlocks metadata-transfer parallel-fetch sandbox sfperms strict"
GENTOO_MIRRORS="ftp://ftp.vic.keypoint.com.au http://mirrors.tds.net/gentoo ftp://mirrors.tds.net/gentoo http://mirror.tucdemonic.org/gentoo/"
LC_ALL="en_GB.UTF-8"
LDFLAGS="-Wl,--as-needed"
LINGUAS="fr en_GB"
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/portage/local/layman/science /usr/portage/local/layman/ebuild-exchange"
SYNC="rsync://rsync.au.gentoo.org/gentoo-portage"
USE="ppc X a52 aac aalib alsa altivec amr arts asf audiofile berkdb bitmap-fonts bonobo bzip2 cairo cddb cdf cdparanoia cdr cli cracklib crypt cups curl divx dlloader dri dts dv dvd dvdr dvdread elibc_glibc emboss encode esd expat f77 ffmepg fftw firefox flac foomaticdb fortran gdbm gif gmedia gmp gpm graphviz gs gstreamer gtk gtk2 gtkhtml hal iconv idn ieee1394 imagemagick imlib imlib2 input_devices_evdev input_devices_keyboard input_devices_mouse ipv6 isdnlog java jbig jpeg jpeg2k kde kdehiddenvisibility kdexdeltas kernel_linux lcms libg++ libwww linguas_en_GB linguas_fr live lzo mad mikmod mjpeg mng mp3 mp4 mpeg musepack musicbrainz ncurses netcdf network nls nptl nptlonly ogg openal openexr opengl oss pam pcre pdf perl png ppds pppd python qt3 qt4 quicktime rdesktop readline reflection ruby samba sdl session slang slp sndfile spell spl ssl svg szip tcl tcltk tcpd tetex theora tiff tk truetype truetype-fonts type1-fonts udev unicode usb userland_GNU vcd video_cards_fbdev video_cards_nv vorbis wmf wmp x264 xine xml xml2 xorg xv xvid zeroconf zlib"
Unset: CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, MAKEOPTS, PORTAGE_RSYNC_EXTRA_OPTS
I didn't use special CFLAGS for this run.
Sorry in my last comment the log was cut at some column, here it in full width: Sun Nov 5 12:30:42 2006 (01:13 min.) ###################################################### Registering python UNO components: ###################################################### SUCCESS: Source for types.rdb: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb SUCCESS: Source for pyuno_services.rdb: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/pyuno_services.rdb Systemcall: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/pyuno_services.rdb -r /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_inprogress_1/services.rdb -c vnd.openoffice.pymodule:mailmerge -l com.sun.star.loader.Python 2>&1 | ERROR: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/pyuno_services.rdb -r /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_inprogress_1/services.rdb -c vnd.openoffice.pymodule:mailmerge -l com.sun.star.loader.Python 2>&1 | Systemcall: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/pyuno_services.rdb -r /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_inprogress_1/services.rdb -c vnd.openoffice.pymodule:pythonscript -l com.sun.star.loader.Python 2>&1 | ERROR: /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/regcomp -register -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/types.rdb -br /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/solver/680/unxlngppc.pro/bin/pyuno_services.rdb -r /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_inprogress_1/services.rdb -c vnd.openoffice.pymodule:pythonscript -l com.sun.star.loader.Python 2>&1 | Moved directory from /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_inprogress_1 to /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_witherror_1 Removing directory /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//zip/fr_en-GB Removing directory /var/tmp/portage/openoffice-2.0.4/work/ooo/build/OOO_2_0_4/instsetoo_native/util/OpenOffice//services.rdb/fr_en-GB_witherror_1 (In reply to comment #48) > I've added a new patch to the ebuild from upstream, which tries to solve the > regcomp segfaults. So could you please try out the new version and see if the > fix helps in your case? Nope, still fails with the same error on a G4 PowerPC. But not on x86 any more (where I also had the pyUNO-bug). On x86 it fails now with chown: changing ownership of `/var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice': No such file or directory # ls -l /var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice lrwxrwxrwx 1 root root 35 4. Nov 18:42 /var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice -> /usr/lib/openoffice/program/soffice And as the target does not exist yet, it fails... (In reply to comment #51) > > Nope, still fails with the same error on a G4 PowerPC. So much for that hope... > But not on x86 any more (where I also had the pyUNO-bug). On x86 it fails now > with > > chown: changing ownership of > `/var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice': No such file or > directory > > # ls -l /var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice > lrwxrwxrwx 1 root root 35 4. Nov 18:42 > /var/tmp/portage/openoffice-2.0.4/image/usr/bin/soffice -> > /usr/lib/openoffice/program/soffice > > And as the target does not exist yet, it fails... > Actually that would be another bug, so maybe you could open a new one, as it works fine here. (also: why didn't you open a report about the pyuno-problem on x86? Cause there it should work fine...) I just started a new installation. I will let you now if it fails or suceeds in few hours :) (In reply to comment #52) > Actually that would be another bug, so maybe you could open a new one, as it > works fine here. See bug #154230 now. > (also: why didn't you open a report about the pyuno-problem on > x86? Cause there it should work fine...) I compiled OpenOffice a couple of times with different CFLAGS and python-settings, just to make sure, that the problem is not on my side. But actually I never commented on that bug here -- no clue why I forgot it. Fails here as before, too. *** Bug 157293 has been marked as a duplicate of this bug. *** CFLAGS="-Os -mcpu=7400 -mtune=7400 -fno-strict-aliasing -pipe" and re-emerging python worked like a charm here. Is this bug the same as http://www.openoffice.org/issues/show_bug.cgi?id=71078 ? If so, it appears that it will be fixed in ooo-2.1. (In reply to comment #58) > Is this bug the same as http://www.openoffice.org/issues/show_bug.cgi?id=71078 > ? > > If so, it appears that it will be fixed in ooo-2.1. > Unfortunately not. We are already using the fix from above, didn't help. @Lars and others: I've just commited some more patches / changes which try to fix the regcomp problem (by disabling it). But as I don't have the hardware couldn't test, so could you please try again (especially as we need it for security bug #159951) (In reply to comment #60) > @Lars and others: I've just commited some more patches / changes which try to > fix the regcomp problem (by disabling it). But as I don't have the hardware > couldn't test, so could you please try again (especially as we need it for > security bug #159951) Success! :-D Mon Jan 8 07:47:48 2007 >>> app-office/openoffice-2.0.4 merge time: 15 hours, 12 minutes and 27 seconds. USE="branding cairo cups dbus eds firefox gnome gstreamer gtk kde ldap sound pam webdav -binfilter -odk -java -debug" CFLAGS="-O2 -mtune=G4 -mcpu=G4 -maltivec -mabi=altivec -pipe -D_STLP_STRICT_ANSI" But it would be good to hear from another ppc-team-member if it installs cleanly as well. (In reply to comment #61) > (In reply to comment #60) > > @Lars and others: I've just commited some more patches / changes which try to > > fix the regcomp problem (by disabling it). But as I don't have the hardware > > couldn't test, so could you please try again (especially as we need it for > > security bug #159951) > > Success! :-D > > Mon Jan 8 07:47:48 2007 >>> app-office/openoffice-2.0.4 > merge time: 15 hours, 12 minutes and 27 seconds. > > USE="branding cairo cups dbus eds firefox gnome gstreamer gtk kde ldap sound > pam webdav -binfilter -odk -java -debug" > CFLAGS="-O2 -mtune=G4 -mcpu=G4 -maltivec -mabi=altivec -pipe > -D_STLP_STRICT_ANSI" > > But it would be good to hear from another ppc-team-member if it installs > cleanly as well. > I'll try on my pegasos ;) (In reply to comment #60) > @Lars and others: I've just commited some more patches / changes which try to > fix the regcomp problem (by disabling it). But as I don't have the hardware > couldn't test, so could you please try again (especially as we need it for > security bug #159951) Worked here, too :) Mac Mini G4 with: CFLAGS="-O2 -mcpu=G4 -mtune=G4 -maltivec -mabi=altivec -fno-strict-aliasing -pipe" and USE="binfilter cairo firefox gtk kde odk" for ooo. And another success with USE="cups dbus eds firefox gnome gstreamer gtk kde ldap pam -binfilter -branding -cairo -debug -java -odk -sound -webdav" and CFLAGS="-O2 -mcpu=G3 -mtune=G3 -fno-strict-aliasing -pipe" which also caused troubles in the past during GRP packages building. I guess this counts as fixed now, very happy to close the bug, finally ;) |