Re: is 'make check' supposed to work on ~master ?
On Fri, Dec 14, 2012 at 8:00 AM, Noel Grandin n...@peralex.com wrote: On 2012-12-13 19:28, John Smith wrote: Conditional jump or move depends on uninitialised value(s) That looks like you have valgrind enabled, and I don't think all of our unit tests have been made valgrind-safe yet. Oh. Well, when I do 'make build' without valgrind, I get a segmentation fault : [build CUT] sw_subsequent_rtfexport [build CUT] sw_subsequent_rtfimport /bin/sh: line 1: 4079 Segmentation fault (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}$O/lib:$O/lib/sqlite DBGSV_ERROR_OUT=shell DISABLE_SAL_DBGBOX=t STAR_RESOURCEPATH=$O/bin/ $O/bin/cppunit/cppunittester $W/LinkTarget/CppunitTest/libtest_sw_subsequent_ww8export.so --headless -env:BRAND_BASE_DIR=file://$O/unittest/install -env:CONFIGURATION_LAYERS=xcsxcu:file://$O/xml/registry module:file://$O/xml/registry/spool xcsxcu:file://$O/unittest/registry -env:UNO_TYPES=file://$O/bin/offapi.rdb file://$O/bin/udkapi.rdb -env:UNO_SERVICES=file://$O/xml/ure/services.rdb file://$O/xml/component/basic/util/sb.component file://$O/xml/component/comphelper/util/comphelp.component file://$O/xml/component/configmgr/source/configmgr.component file://$O/xml/component/dbaccess/util/dba.component file://$O/xml/component/embeddedobj/util/embobj.component file://$O/xml/component/fileaccess/source/fileacc.component file://$O/xml/component/filter/source/config/cache/filterconfig1.component file://$O/xml/component/forms/util/frm.component file://$O/xml/component/framework/util/fwk.component file://$O/xml/component/i18npool/util/i18npool.component file://$O/xml/component/linguistic/source/lng.component file://$O/xml/component/package/source/xstor/xstor.component file://$O/xml/component/package/util/package2.component file://$O/xml/component/sax/source/expatwrap/expwrap.component file://$O/xml/component/sw/util/msword.component file://$O/xml/component/sw/util/sw.component file://$O/xml/component/sw/util/swd.component file://$O/xml/component/sfx2/util/sfx.component file://$O/xml/component/svl/source/fsstor/fsstorage.component file://$O/xml/component/svtools/util/svt.component file://$O/xml/component/toolkit/util/tk.component file://$O/xml/component/ucb/source/core/ucb1.component file://$O/xml/component/ucb/source/ucp/file/ucpfile1.component file://$O/xml/component/unotools/util/utl.component file://$O/xml/component/unoxml/source/service/unoxml.component file://$O/xml/component/xmlhelp/util/ucpchelp1.component -env:URE_INTERNAL_LIB_DIR=file://$O/lib -env:LO_LIB_DIR=file://$O/lib --protector unoexceptionprotector.so unoexceptionprotector --protector unobootstrapprotector.so unobootstrapprotector $W/CppunitTest/sw_subsequent_ww8export.test.log 21 OK (1) Error: a unit test failed, please do one of: export DEBUGCPPUNIT=TRUE# for exception catching export GDBCPPUNITTRACE=gdb --args # for interactive debugging export VALGRIND=memcheck# for memory checking and retry. make[2]: *** [/usr/local/src/libreoffice/workdir/unxlngi6.pro/CppunitTest/sw_subsequent_ww8export.test] Error 1 make[2]: *** Waiting for unfinished jobs --- Oh dear - something failed during the build - sorry ! For more help with debugging build errors, please see the section in: http://wiki.documentfoundation.org/Development internal build errors: ERROR: error 512 occurred while making /usr/local/src/libreoffice/tail_build/prj it seems that the error is inside 'tail_build', please re-run build inside this module to isolate the error and/or test your fix. --- To rebuild a specific module: make tail_build.clean # not recommended, this will re-build almost everything make tail_build when the problem is isolated and fixed, re-run 'make' make[1]: *** [build-packimages] Error 1 make[1]: Leaving directory `/usr/local/src/libreoffice' make: *** [build] Error 2 [root@localhost libreoffice]# gdb /usr/local/src/libreoffice/solver/unxlngi6.pro/bin/cppunit/cppunittester ./tail_build/core.4079 GNU gdb (GDB) Fedora (7.4.50.20120120-49.fc17) Copyright (C) 2012 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type show copying and show warranty for details. This GDB was configured as i686-redhat-linux-gnu. For bug reporting instructions, please see: http://www.gnu.org/software/gdb/bugs/... Reading symbols from /usr/local/src/libreoffice/solver/unxlngi6.pro/bin/cppunit/cppunittester...(no debugging symbols found)...done. [New LWP 4079] [New LWP 4080] warning: the debug information found in /usr/lib/debug/usr/lib/libstdc++.so.6.0.17.debug does not match /lib/libstdc++.so.6 (CRC mismatch). warning: the debug information found in
Re: is 'make check' supposed to work on ~master ?
On 12/12/2012 06:36 PM, John Smith wrote: Im trying to run 'make check' at toplevel on ~master. It fails for me at this point (soffice.bin crashes) : make /usr/local/src/libreoffice/workdir/unxlngi6.pro/JunitTest/comphelper_complex/done Is that crash reproducible for you? The backtrace doesn't give much of a clue to why it fails, so if it is reproducible running under Valgrind might help. In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), so I recommend to configure with --disable-online-update (which is the implicit default on Linux anyway, though). But the backtrace here doesn't look like it is caused by that problem anyway. Stephan ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
On 2012-12-13 11:35, Stephan Bergmann wrote: In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), Is it not possible to pass in a parameter disabling the online-update-check? That should make the tests happier. Disclaimer: http://www.peralex.com/disclaimer.html ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
On 12/13/2012 10:37 AM, Noel Grandin wrote: On 2012-12-13 11:35, Stephan Bergmann wrote: In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), Is it not possible to pass in a parameter disabling the online-update-check? That should make the tests happier. I have the faint memory that we indeed somehow disabled it at least for some of the tests in the past, but can't find any relevant code. (There's apparently no such command-line parameter for soffice, and the code at extensions/source/update/check/updatecheckjob.cxx also does not check any environment variable or similar.) Stephan ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
On 2012-12-13 11:53, Stephan Bergmann wrote: On 12/13/2012 10:37 AM, Noel Grandin wrote: On 2012-12-13 11:35, Stephan Bergmann wrote: In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), Is it not possible to pass in a parameter disabling the online-update-check? That should make the tests happier. I have the faint memory that we indeed somehow disabled it at least for some of the tests in the past, but can't find any relevant code. (There's apparently no such command-line parameter for soffice, and the code at extensions/source/update/check/updatecheckjob.cxx also does not check any environment variable or similar.) Sounds like something I could log as a MediumHack? Disclaimer: http://www.peralex.com/disclaimer.html ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
On Thu, Dec 13, 2012 at 10:35 AM, Stephan Bergmann sberg...@redhat.com wrote: On 12/12/2012 06:36 PM, John Smith wrote: Im trying to run 'make check' at toplevel on ~master. It fails for me at this point (soffice.bin crashes) : make /usr/local/src/libreoffice/workdir/unxlngi6.pro/JunitTest/comphelper_complex/done Is that crash reproducible for you? The backtrace doesn't give much of a clue to why it fails, so if it is reproducible running under Valgrind might help. In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), so I recommend to configure with --disable-online-update (which is the implicit default on Linux anyway, though). But the backtrace here doesn't look like it is caused by that problem anyway. It's pretty much reproducable for me: I re-tried make a few times. Ill retry with ./configure --disable-online-update make build make check, and see if it still fails. - John Smith ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
On 12/13/2012 11:00 AM, Noel Grandin wrote: On 2012-12-13 11:53, Stephan Bergmann wrote: On 12/13/2012 10:37 AM, Noel Grandin wrote: On 2012-12-13 11:35, Stephan Bergmann wrote: In general, the online update mechanism spawns threads that it doesn't properly join again before exit, which is a notorious problem with tests (that often run quickly enough to terminate soffice.bin while the online update check is still in progress), Is it not possible to pass in a parameter disabling the online-update-check? That should make the tests happier. I have the faint memory that we indeed somehow disabled it at least for some of the tests in the past, but can't find any relevant code. (There's apparently no such command-line parameter for soffice, and the code at extensions/source/update/check/updatecheckjob.cxx also does not check any environment variable or similar.) Sounds like something I could log as a MediumHack? Preferred hack would of course be to fix that online update code, but it looks like an opaque disaster at least to me... Stephan ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: is 'make check' supposed to work on ~master ?
Well, I re-tried 'make build ; make check' with './configure --disable-online-update'. But I run into the the following : (Maybe I should leave this alone for now ...) [build CUT] sw_subsequent_ooxmlimport ods Test xls Test xlsx Test csv Test ==5755== Conditional jump or move depends on uninitialised value(s) ==5755==at 0xE43D5E3: std::_Rb_treeunsigned long, std::pairunsigned long const, SvNumberformat*, std::_Select1ststd::pairunsigned long const, SvNumberformat* , std::lessunsigned long, std::allocatorstd::pairunsigned long const, SvNumberformat* ::_M_lower_bound(std::_Rb_tree_nodestd::pairunsigned long const, SvNumberformat* *, std::_Rb_tree_nodestd::pairunsigned long const, SvNumberformat* *, unsigned long const) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so) ==5755==by 0xE437F69: std::_Rb_treeunsigned long, std::pairunsigned long const, SvNumberformat*, std::_Select1ststd::pairunsigned long const, SvNumberformat* , std::lessunsigned long, std::allocatorstd::pairunsigned long const, SvNumberformat* ::find(unsigned long const) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so) ==5755==by 0xE432D2A: std::mapunsigned long, SvNumberformat*, std::lessunsigned long, std::allocatorstd::pairunsigned long const, SvNumberformat* ::find(unsigned long const) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so) ==5755==by 0xE4165E7: SvNumberFormatter::GetFormatEntry(unsigned long) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so) ==5755==by 0xE40BD31: SvNumberFormatter::IsNumberFormat(rtl::OUString const, unsigned long, double) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so) ==5755==by 0xA532CC0: ScColumn::SetString(long, short, String const, formula::FormulaGrammar::AddressConvention, ScSetStringParam*) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xA967AD1: ScTable::SetString(short, long, short, String const, ScSetStringParam*) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xA67C26C: ScDocument::SetString(short, long, short, rtl::OUString const, ScSetStringParam*) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xB45C020: lcl_PutString(ScDocument*, short, long, short, String const, unsigned char, SvNumberFormatter*, bool, utl::TransliterationWrapper, CalendarWrapper, utl::TransliterationWrapper*, CalendarWrapper*) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xB45EE90: ScImportExport::ExtText2Doc(SvStream) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xB4544D3: ScImportExport::ImportStream(SvStream, String const, unsigned long) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xB3415CB: ScDocShell::ConvertFrom(SfxMedium) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsclo.so) ==5755==by 0xD8F5355: SfxObjectShell::DoLoad(SfxMedium*) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsfxlo.so) ==5755==by 0x65EF318: ScFiltersTest::load(rtl::OUString const, rtl::OUString const, rtl::OUString const, rtl::OUString const, unsigned int, unsigned int, unsigned int) (in /usr/local/src/libreoffice/workdir/unxlngi6.pro/LinkTarget/CppunitTest/libtest_sc_subsequent_filters_test.so) ==5755==by 0x662804E: ScFiltersTest::testBrokenQuotesCSV() (in /usr/local/src/libreoffice/workdir/unxlngi6.pro/LinkTarget/CppunitTest/libtest_sc_subsequent_filters_test.so) ==5755==by 0x666F4D8: CppUnit::TestCallerScFiltersTest::runTest() (in /usr/local/src/libreoffice/workdir/unxlngi6.pro/LinkTarget/CppunitTest/libtest_sc_subsequent_filters_test.so) ==5755==by 0x4B895E10: CppUnit::TestCaseMethodFunctor::operator()() const (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4C337CD: (anonymous namespace)::Prot::protect(CppUnit::Functor const, CppUnit::ProtectorContext const) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/unobootstrapprotector.so) ==5755==by 0x4B8930F1: CppUnit::ProtectorChain::ProtectFunctor::operator()() const (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4012ED5: (anonymous namespace)::Prot::protect(CppUnit::Functor const, CppUnit::ProtectorContext const) (in /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/unoexceptionprotector.so) ==5755==by 0x4B8930F1: CppUnit::ProtectorChain::ProtectFunctor::operator()() const (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4B88BDB9: CppUnit::DefaultProtector::protect(CppUnit::Functor const, CppUnit::ProtectorContext const) (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4B8930F1: CppUnit::ProtectorChain::ProtectFunctor::operator()() const (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4B892DC8: CppUnit::ProtectorChain::protect(CppUnit::Functor const, CppUnit::ProtectorContext const) (in /usr/lib/libcppunit-1.12.so.1.0.0) ==5755==by 0x4B89C5DD:
Re: is 'make check' supposed to work on ~master ?
On 2012-12-13 19:28, John Smith wrote: Conditional jump or move depends on uninitialised value(s) That looks like you have valgrind enabled, and I don't think all of our unit tests have been made valgrind-safe yet. Disclaimer: http://www.peralex.com/disclaimer.html ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
is 'make check' supposed to work on ~master ?
Hi, Im trying to run 'make check' at toplevel on ~master. It fails for me at this point (soffice.bin crashes) : make /usr/local/src/libreoffice/workdir/unxlngi6.pro/JunitTest/comphelper_complex/done Is that 'supposed to work', or should I expect failure(s) ? - John Smith -- stacktrace -- It seems like soffice.bin crashed during the test excution! Found a core dump at /usr/local/src/libreoffice/workdir/unxlngi6.pro/JunitTest/comphelper_complex/user/core.23956 Stacktrace: [New LWP 23956] [New LWP 23957] [New LWP 23960] warning: the debug information found in /usr/lib/debug/usr/lib/libstdc++.so.6.0.17.debug does not match /lib/libstdc++.so.6 (CRC mismatch). warning: the debug information found in /usr/lib/debug//usr/lib/libstdc++.so.6.0.17.debug does not match /lib/libstdc++.so.6 (CRC mismatch). warning: the debug information found in /usr/lib/debug/usr/lib//libstdc++.so.6.0.17.debug does not match /lib/libstdc++.so.6 (CRC mismatch). [Thread debugging using libthread_db enabled] Using host libthread_db library /lib/libthread_db.so.1. Core was generated by `/usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/soffice'. Program terminated with signal 11, Segmentation fault. #0 0x46ebbcab in _int_free (av=av@entry=0x46ff2420, p=0x9949118, have_lock=have_lock@entry=1) at malloc.c:4103 4103unlink(av, nextchunk, bck, fwd); warning: File /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libuno_sal.so.3-gdb.py auto-loading has been declined by your `auto-load safe-path' set to $debugdir:$datadir/auto-load:/usr/bin/mono-gdb.py. warning: File /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libuno_cppu.so.3-gdb.py auto-loading has been declined by your `auto-load safe-path' set to $debugdir:$datadir/auto-load:/usr/bin/mono-gdb.py. warning: File /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libsvllo.so-gdb.py auto-loading has been declined by your `auto-load safe-path' set to $debugdir:$datadir/auto-load:/usr/bin/mono-gdb.py. warning: File /usr/local/src/libreoffice/solver/unxlngi6.pro/lib/libtllo.so-gdb.py auto-loading has been declined by your `auto-load safe-path' set to $debugdir:$datadir/auto-load:/usr/bin/mono-gdb.py. Thread 3 (Thread 0xadc6cb40 (LWP 23960)): #0 0xb77ea424 in __kernel_vsyscall () #1 0x46f3a968 in accept () at ../sysdeps/unix/sysv/linux/i386/socket.S:97 #2 0xb76e8b46 in osl_acceptPipe () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #3 0xb75c6492 in osl::Pipe::accept(osl::StreamPipe) () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/libsofficeapp.so #4 0xb75beccb in desktop::OfficeIPCThread::execute() () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/libsofficeapp.so #5 0xb6a99a2f in salhelper::Thread::run() () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_salhelpergcc3.so.3 #6 0xb6a9a21a in threadFunc () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_salhelpergcc3.so.3 #7 0xb76f8525 in osl_thread_start_Impl () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #8 0x46ffeadf in start_thread (arg=0xadc6cb40) at pthread_create.c:309 #9 0x46f3954e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:133 Thread 2 (Thread 0xb00fab40 (LWP 23957)): #0 0xb77ea424 in __kernel_vsyscall () #1 0x470024d4 in pthread_cond_timedwait@@GLIBC_2.3.2 () at ../nptl/sysdeps/unix/sysv/linux/i386/i486/pthread_cond_timedwait.S:238 #2 0x46f479a4 in __pthread_cond_timedwait (cond=0xb77cdaa0, mutex=0xb77cd940, abstime=0xb00f72e0) at forward.c:152 #3 0xb7711090 in rtl_cache_wsupdate_wait(unsigned int) () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #4 0xb771159d in rtl_cache_wsupdate_all(void*) () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #5 0x46ffeadf in start_thread (arg=0xb00fab40) at pthread_create.c:309 #6 0x46f3954e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:133 Thread 1 (Thread 0xb02ff700 (LWP 23956)): #0 0x46ebbcab in _int_free (av=av@entry=0x46ff2420, p=0x9949118, have_lock=have_lock@entry=1) at malloc.c:4103 #1 0x46ebe180 in free_check (mem=0x9949180, caller=0xb77132f7) at hooks.c:259 #2 0x46ebf9e1 in __GI___libc_free (mem=0x9949180) at malloc.c:2964 #3 0xb77132f7 in rtl_freeMemory_SYSTEM(void*) () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #4 0xb771361e in rtl_freeMemory () from /usr/local/src/libreoffice/solver/unxlngi6.pro/installation/opt/program/../ure-link/lib/libuno_sal.so.3 #5 0xb6c276c0 in cppu::OWeakObject::operator delete(void*) () from