Re: [Libreoffice] segfault in libqa_unit.so test
On Mon, 2011-01-10 at 06:50 +0100, David Tardon wrote: > > - CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) > > $(SOLARBINDIR)/cppunittester > > + CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) gdb > > --args $(SOLARBINDIR)/cppunittester > Actually, no tweaking is necessary, as it is already in place :) Just > use > > dmake debugcppunit=t Well, that one tries to log all try/catches and get a bt for all of them so manually tweaking it is still probably easier than dealing with the -nx gdb used in that mode. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Fri, Jan 07, 2011 at 09:52:31AM +, Caolán McNamara wrote: > On Fri, 2011-01-07 at 03:21 +0100, Thomas Klausner wrote: > > Is that the end of the build? > > Yup, good news. > > > (Where did the nice graphics go? ;) ) > > > > Thank you! How do we continue with this? > > Well, make dev-install and see if it runs :-) > > The _exit is of course a temporary hack, but it tests the theory that > your crashes were related to post-main problems, probably some global > object dtor problems of some kind. So we really need to isolate that > better, e.g. removing the _exit hack and tweak solenv/inc/settings.mk > > - CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) > $(SOLARBINDIR)/cppunittester > + CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) gdb > --args $(SOLARBINDIR)/cppunittester Actually, no tweaking is necessary, as it is already in place :) Just use dmake debugcppunit=t > > and re-run build in whatever module was originally crashing and get a > backtrace to see if we can get a hint as to what the crasher might be. > > C. > > ___ > LibreOffice mailing list > LibreOffice@lists.freedesktop.org > http://lists.freedesktop.org/mailman/listinfo/libreoffice ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Fri, Jan 07, 2011 at 03:21:06AM +0100, Thomas Klausner wrote: > Multiprocessing build is finished > Maximal number of processes run: 1 > > Is that the end of the build? (Where did the nice graphics go? ;) ) Yes - the asciiart is part of build.git - it could be ported here. ;) pgpU6xS5BdCgz.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
Hi Thomas, On Fri, 2011-01-07 at 03:21 +0100, Thomas Klausner wrote: > With a new pull and your patch applied, I get to: > No EPM: do no packaging at this stage Looks good to me :-) > Multiprocessing build is finished > Maximal number of processes run: 1 > > Is that the end of the build? (Where did the nice graphics go? ;) ) This is a really good question :-) I think we lost them when we migrated from build/ - it would be great to resurrect that fun and re-commit it. > Thank you! How do we continue with this? After it is built: 'ooinstall' (with shell environment sourced) is your friend. HTH, Michael. -- michael.me...@novell.com <><, Pseudo Engineer, itinerant idiot ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Fri, 2011-01-07 at 03:21 +0100, Thomas Klausner wrote: > Is that the end of the build? Yup, good news. > (Where did the nice graphics go? ;) ) > > Thank you! How do we continue with this? Well, make dev-install and see if it runs :-) The _exit is of course a temporary hack, but it tests the theory that your crashes were related to post-main problems, probably some global object dtor problems of some kind. So we really need to isolate that better, e.g. removing the _exit hack and tweak solenv/inc/settings.mk - CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) $(SOLARBINDIR)/cppunittester + CPPUNITTESTER=$(AUGMENT_LIBRARY_PATH_LOCAL) $(GDBCPPUNITTRACE) gdb --args $(SOLARBINDIR)/cppunittester and re-run build in whatever module was originally crashing and get a backtrace to see if we can get a hint as to what the crasher might be. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Thu, Jan 06, 2011 at 10:18:38PM +, Caolán McNamara wrote: > Bah, let me rethink that. I guess its very possible that _exit doesn't > flush standard I/O on NetBSD which might indeed cause an empty > generated .hxx. I've just pushed a fflush into that generator to help > test the theory. Try a pull and build again of tools. With a new pull and your patch applied, I get to: No EPM: do no packaging at this stage Multiprocessing build is finished Maximal number of processes run: 1 Is that the end of the build? (Where did the nice graphics go? ;) ) Thank you! How do we continue with this? Thomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Thu, 2011-01-06 at 22:07 +, Caolán McNamara wrote: > On Thu, 2011-01-06 at 22:44 +0100, Thomas Klausner wrote: > > On Tue, Jan 04, 2011 at 03:01:53PM +, Caolán McNamara wrote: > > > One horrific temporary hack to test a theory would be to call _exit to > > > skip calling the massive dtor chain which OOo builds up. Though that's > > > not a cure, just a temp hack, but maybe it'll help isolate this. > > > > I'm confused why, but with this patch applied, I get failure much > > earlier. > > Unrelated I feel. Bah, let me rethink that. I guess its very possible that _exit doesn't flush standard I/O on NetBSD which might indeed cause an empty generated .hxx. I've just pushed a fflush into that generator to help test the theory. Try a pull and build again of tools. reversemap.obj depends on reversemap.hxx which depends on "bestreversemap" which I've added the extra fflush to so the rebuild should re-generate it without any manual intervention. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Thu, 2011-01-06 at 22:44 +0100, Thomas Klausner wrote: > On Tue, Jan 04, 2011 at 03:01:53PM +, Caolán McNamara wrote: > > One horrific temporary hack to test a theory would be to call _exit to > > skip calling the massive dtor chain which OOo builds up. Though that's > > not a cure, just a temp hack, but maybe it'll help isolate this. > > I'm confused why, but with this patch applied, I get failure much > earlier. Unrelated I feel. > > Making:test_tools.so > ../unxbsdx3.pro/slo/test_reversemap.o: In function `(anonymous > namespace)::Test::testEncoding(unsigned short)': > test_reversemap.cxx:(.text+0x3f0): undefined reference to > `getBestMSEncodingByChar(unsigned short)' > dmake: Error code 1, while making '../unxbsdx3.pro/lib/test_tools.so' > Forcing regeneration of dependency info export VERBOSE=true here to get the command lines used. fiddle with the equivalent of nm -D ../unxbsdx3.pro/lib/libtl*so | c++filt|grep getBest to see if the symbol in in libtl*.so like it should be to help find out if we're linking against the right libtl locally e.g. as opposed to the one inside the solver. If its not then see if ../unxbsdx3.pro/bin/bestreversemap exists, and see if ../unxbsdx3.pro/inc/reversemap.hxx exists with getBestMSEncodingByChar inside it. That reversemap.hxx should be generated by bestreversemap and included by source/string/reversemap.cxx and built into the libtl.so My *guess* is that for some reason ../unxbsdx3.pro/inc/reversemap.hxx exists but is empty. Is that the case ? C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Tue, Jan 04, 2011 at 03:01:53PM +, Caolán McNamara wrote: > One horrific temporary hack to test a theory would be to call _exit to > skip calling the massive dtor chain which OOo builds up. Though that's > not a cure, just a temp hack, but maybe it'll help isolate this. I'm confused why, but with this patch applied, I get failure much earlier. Module 'odk' delivered successfully. 3 files copied, 1 files unchanged --- 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 65280 occurred while making /disk/2/archive/libreoffice/tools/qa it seems that the error is inside 'tools', please re-run build inside this module to isolate the error and/or test your fix: --- rm -Rf /disk/2/archive/libreoffice/tools/unxbsdx3.pro # optional module 'clean' /bin/bash cd /disk/2/archive/libreoffice source ./NetBSDX86-64Env.Set.sh cd tools build when the problem is isolated and fixed exit and re-run 'make' from the top-level when I do that: Making:test_tools.so ../unxbsdx3.pro/slo/test_reversemap.o: In function `(anonymous namespace)::Test::testEncoding(unsigned short)': test_reversemap.cxx:(.text+0x3f0): undefined reference to `getBestMSEncodingByChar(unsigned short)' dmake: Error code 1, while making '../unxbsdx3.pro/lib/test_tools.so' Forcing regeneration of dependency info -- - start unit test #1 on library -- : && LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/libs-gui/tools/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester Usage: cppunittester dmake: Error code 1, while making 'test1' Retrying /disk/2/archive/libreoffice/tools/qa Making:all_qa.dpslo Making:test_tools.so ../unxbsdx3.pro/slo/test_reversemap.o: In function `(anonymous namespace)::Test::testEncoding(unsigned short)': test_reversemap.cxx:(.text+0x3f0): undefined reference to `getBestMSEncodingByChar(unsigned short)' dmake: Error code 1, while making '../unxbsdx3.pro/lib/test_tools.so' --- Oh dear - something failed during the build - sorry ! For more help with debugging build errors, please see the section in: http://wiki.documentfoundation.org/Development it seems that the error is inside 'tools', please re-run build inside this module to isolate the error and/or test your fix: --- rm -Rf /disk/2/archive/libreoffice/tools/unxbsdx3.pro # optional module 'clean' /bin/bash cd /disk/2/archive/libreoffice source ./NetBSDX86-64Env.Set.sh cd tools build How can the diff cause that? Thomas ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
On Thu, 2010-12-16 at 12:57 +0100, Thomas Klausner wrote: > The build on NetBSD currently fails with: > > Entering /disk/2/archive/libreoffice/sc/qa/unit > > -- > - start unit test #1 on library ../../unxbsdx3.pro/lib/libqa_unit.so > -- > : && > LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib > /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester > ../../unxbsdx3.pro/lib/libqa_unit.so -headless -invisible \ > > -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb > \ > > -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb > > file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" > \ > > -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" > \ > > -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" > \ > > -env:UNO_USER_PACKAGES_CACHE="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" > OK (3) > /bin/bash: line 1: 21530 Segmentation fault (core dumped) > LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib > /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester > ../../unxbsdx3.pro/lib/libqa_unit.so -headless -invisible > -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb > > -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb > > file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" > > -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" > > -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" > -env:UNO_USER_PACKAGES_CACHE="file:///disk/2/archive/libreoffice/clone/! cal > c/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" Is this still broken for you. Looking at this, it appears to be a crash on exit seeing as it was able to print the "tests passed" line of OK. One horrific temporary hack to test a theory would be to call _exit to skip calling the massive dtor chain which OOo builds up. Though that's not a cure, just a temp hack, but maybe it'll help isolate this. C. diff --git a/sal/inc/sal/main.h b/sal/inc/sal/main.h index bf7a828..65ead5d 100644 --- a/sal/inc/sal/main.h +++ b/sal/inc/sal/main.h @@ -41,6 +41,19 @@ extern "C" { void SAL_CALL sal_detail_initialize(int argc, char ** argv); void SAL_CALL sal_detail_deinitialize(); +#if defined(AIX) || defined(NETBSD) +//*Hurl*, *Puke*, temp fix only, will +//cause other weirdness eventually +# define BROKEN_SHUTDOWN +# include +#endif + +#ifdef BROKEN_SHUTDOWN +# define MAINRET(x) _exit(x) +#else +# define MAINRET(x) return x +#endif + #define SAL_MAIN_WITH_ARGS_IMPL \ int SAL_CALL main(int argc, char ** argv) \ { \ @@ -48,7 +61,7 @@ int SAL_CALL main(int argc, char ** argv) \ sal_detail_initialize(argc, argv); \ ret = sal_main_with_args(argc, argv); \ sal_detail_deinitialize(); \ -return ret; \ +MAINRET(ret); \ } #define SAL_MAIN_IMPL \ @@ -58,7 +71,7 @@ int SAL_CALL main(int argc, char ** argv) \ sal_detail_initialize(argc, argv); \ ret = sal_main(); \ sal_detail_deinitialize(); \ -return ret; \ +MAINRET(ret); \ } /* Definition macros for CRT entries */ ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
Michael Meeks wrote: > > -- > > - start unit test #1 on library ../../unxbsdx3.pro/lib/libqa_unit.so > > -- > > So - I would cut/paste the command-line; export the LD_LIBRARY_PATH > yourself; and run the cppunittester inside gdb to get a trace. Failing > that, run it in valgrind (might help), then you'll prolly need to > re-compile the failing piece with debug symbols enabled to make sense of > the trace: though 'build debug=true' may print more useful output anyway > before failing. > Chiming in here to advertise a bit what Caolan added recently: build debugcppunit=t outputs a proper backtrace, to get a first idea what to compile with debug symbols; might be a good thing to add something similar for valgrind - or is there already? Cheers, -- Thorsten pgpjF3oRhmcgn.pgp Description: PGP signature ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [Libreoffice] segfault in libqa_unit.so test
Hi Thomas, On Thu, 2010-12-16 at 12:57 +0100, Thomas Klausner wrote: > The build on NetBSD currently fails with: > Entering /disk/2/archive/libreoffice/sc/qa/unit This is where we run our unit testing framework, which involves the first real(ish) run of great chunks of the code. > -- > - start unit test #1 on library ../../unxbsdx3.pro/lib/libqa_unit.so > -- So - I would cut/paste the command-line; export the LD_LIBRARY_PATH yourself; and run the cppunittester inside gdb to get a trace. Failing that, run it in valgrind (might help), then you'll prolly need to re-compile the failing piece with debug symbols enabled to make sense of the trace: though 'build debug=true' may print more useful output anyway before failing. HTH, Michael. -- michael.me...@novell.com <><, Pseudo Engineer, itinerant idiot ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
[Libreoffice] segfault in libqa_unit.so test
Hi! I've switched back to the internal odbc library, since the external one gave me the BOOL errors. Strangely, "--with-system-odbc-headers=no" gives me an error by the configure script because it's not recognized, but it makes the configure script switch to the internal headers. Sebastian, want to take a look at that? The build on NetBSD currently fails with: Entering /disk/2/archive/libreoffice/sc/qa/unit -- - start unit test #1 on library ../../unxbsdx3.pro/lib/libqa_unit.so -- : && LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester ../../unxbsdx3.pro/lib/libqa_unit.so -headless -invisible \ -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb \ -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" \ -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" \ -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" \ -env:UNO_USER_PACKAGES_CACHE="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" OK (3) /bin/bash: line 1: 21530 Segmentation fault (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester ../../unxbsdx3.pro/lib/libqa_unit.so -headless -invisible -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" -env:UNO_USER_PACKAGES_CACHE="file:///disk/2/archive/libreoffice/clone/cal c/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" dmake: Error code 139, while making 'test' Forcing regeneration of dependency info nothing to do here... -- - start unit test #1 on library -- : && LD_LIBRARY_PATH=${LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester -headless -invisible \ -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb \ -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" \ -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" \ -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" \ -env:UNO_USER_PACKAGES_CACHE="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" terminate called after throwing an instance of 'CppUnit::DynamicLibraryManagerException' what(): Failed to load dynamic library: -headless /bin/bash: line 1: 28544 Abort trap (core dumped) LD_LIBRARY_PATH=$ {LD_LIBRARY_PATH+${LD_LIBRARY_PATH}:}/disk/2/archive/libreoffice/clone/calc/sc/unxbsdx3.pro/lib:/disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/lib /disk/2/archive/libreoffice/solver/330/unxbsdx3.pro/bin/cppunittester -headless -invisible -env:UNO_SERVICES=file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/services.rdb -env:UNO_TYPES="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/types.rdb file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit/udkapi.rdb" -env:OOO_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/misc/qa_unit" -env:BRAND_BASE_DIR="file:///disk/2/archive/libreoffice/clone/calc/sc/qa/unit/../../unxbsdx3.pro/m