On Mon, Feb 18, 2013 at 12:26 PM, Chris Rees <cr...@freebsd.org> wrote: > On 14 February 2013 13:57, Mikhail T. <mi+t...@aldan.algebra.com> wrote: >> Hello! >> >> I just finished building editors/libreoffice with gcc-4.2.1 -- had to >> edit the port's Makefile to prevent it from picking a different >> compiler. Everything built and installed, but libreoffice dies on >> start-up (right after flashing the splash-window): >> >> (gdb) where >> #0 0x000000080596c1aa in cppu::__getTypeEntries () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #1 0x000000080596c333 in cppu::__queryDeepNoXInterface () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #2 0x000000080596d4a2 in cppu::WeakImplHelper_query () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #3 0x00000008116f2b03 in >> >> cppu::WeakImplHelper1<com::sun::star::lang::XEventListener>::queryInterface >> () >> from /opt/lib/libreoffice/ure/lib/bootstrap.uno.so >> #4 0x0000000805970347 in >> cppu::OInterfaceContainerHelper::disposeAndClear () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #5 0x00000008059705b2 in >> cppu::OMultiTypeInterfaceContainerHelper::disposeAndClear () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #6 0x000000080593309f in cppu::OComponentHelper::dispose () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #7 0x0000000805963d00 in cppu::OFactoryComponentHelper::dispose () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #8 0x00000008116ec296 in stoc_smgr::OServiceManager::disposing () >> from /opt/lib/libreoffice/ure/lib/bootstrap.uno.so >> #9 0x000000080596af05 in cppu::WeakComponentImplHelperBase::dispose () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #10 0x00000008116e6244 in stoc_smgr::ORegistryServiceManager::dispose () >> from /opt/lib/libreoffice/ure/lib/bootstrap.uno.so >> #11 0x000000080596a573 in cppu::WeakComponentImplHelperBase::release () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #12 0x00000008059482f6 in (anonymous namespace)::createTypeRegistry () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #13 0x00000008059487bf in >> cppu::defaultBootstrap_InitialComponentContext () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #14 0x0000000805948918 in >> cppu::defaultBootstrap_InitialComponentContext () >> from >> /opt/lib/libreoffice/program/../ure-link/lib/libuno_cppuhelpergcc3.so.3 >> #15 0x000000080212f883 in >> desktop::Desktop::InitApplicationServiceManager () >> from /opt/lib/libreoffice/program/libmergedlo.so >> #16 0x000000080211f362 in desktop::Desktop::Init () from >> /opt/lib/libreoffice/program/libmergedlo.so >> #17 0x0000000807622113 in InitVCL () from >> /opt/lib/libreoffice/program/libvcllo.so >> #18 0x0000000807623151 in ImplSVMain () from >> /opt/lib/libreoffice/program/libvcllo.so >> #19 0x00000008076232d5 in SVMain () from >> /opt/lib/libreoffice/program/libvcllo.so >> #20 0x000000080214942e in soffice_main () from >> /opt/lib/libreoffice/program/libmergedlo.so >> #21 0x0000000000400773 in main () >> >> I do not blame the office@ team -- the port did not want to use >> gcc-4.2.1, I forced it to. But I'd like to know, what is wrong with the >> compiler shipped by FreeBSD-9.1 (and the only one, if WITHOUT_CLANG is >> defined), that prevents building a healthy libreoffice? >> >> Is there a bug fixed in gcc-4.6? Or is it some (incorrect) assumption >> made by libreoffice code? Thank you, > > Hi Mikhail, > > Libreoffice and openoffice have traditionally recommended that one use > binary packages instead of building it from scratch. > > I'm sure you understand that our compiler in base is rather elderly, > and that a project as insanely huge as Libreoffice is going to be > highly sensitive to minute changes. As a consequence, some very > narrow criteria are chosen to make maintenance of the port possible. > > You are welcome to try with gcc-4.6, but the last I heard it will only > build with clang. Your mileage may vary, please let us know of > success stories!
Just for the record, is find that it works fine for me with gcc-4.6. 9.1-STABLE on i386 system. Building it with the default compiler results in a successful build, but the program would simply exit after a few seconds with no error. The exist status was 0. No messages. When I built with 4.6, it builds and runs fine, at least for the things I've tried. (4.6 invoked by setting WITH_GCC.) -- R. Kevin Oberman, Network Engineer E-mail: rkober...@gmail.com _______________________________________________ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
