RE: qa-build errros on clean code
Jonathan Aquilina wrote (19 februari 2012 17:31) >What version of make does your system have? It seems like the 3.81 >version has a bug with multithreaded builds. I know there is a patch >that was compiled by one of the devs, just not sure where you can get it >the patch for it. 3.82-140.1 Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: qa-build errros on clean code
On 19/02/2012 10:14, Winfried Donkers wrote: Markus Mohrhard wrote (15 februari 2012 19:05) I don't like this solution but well the other option is to debug ucalc and find why it fails ( it does not for me ). Since you now disabled the test can you at least check after a full build that if you start calc insert a note in B2 and then insert a column before B2 the note is moved to C2. Ideally you would even check that inserting a row will move the note to C3. Presently, the problem mentioned earlier doesn't occur :) I did however get a segmentation error once or twice on builds, whichj make me suspect that number of processes/number of cpus may have something to do with it. I am absolutely no expert on this, but I hope this information helps you. Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice What version of make does your system have? It seems like the 3.81 version has a bug with multithreaded builds. I know there is a patch that was compiled by one of the devs, just not sure where you can get it the patch for it. Regards Jonathan ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
RE: qa-build errros on clean code
Markus Mohrhard wrote (15 februari 2012 19:05) >I don't like this solution but well the other option is to debug ucalc >and find why it fails ( it does not for me ). Since you now disabled >the test can you at least check after a full build that if you start >calc insert a note in B2 and then insert a column before B2 the note >is moved to C2. Ideally you would even check that inserting a row will >move the note to C3. Presently, the problem mentioned earlier doesn't occur :) I did however get a segmentation error once or twice on builds, whichj make me suspect that number of processes/number of cpus may have something to do with it. I am absolutely no expert on this, but I hope this information helps you. Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
RE: qa-build errros on clean code
Markus Mohrhard wrote (15 februari 2012 19:05) >I don't like this solution but well the other option is to debug ucalc >and find why it fails ( it does not for me ). Since you now disabled >the test can you at least check after a full build that if you start >calc insert a note in B2 and then insert a column before B2 the note >is moved to C2. Ideally you would even check that inserting a row will >move the note to C3. I wouldn't know how to debug ucalc as the error occurs during make. If you tell me how I am willing to help, but I may have limited time. I took a full build (after git pull -r and disabling the ucalc-lines that produce the assertion), started calc and did what you asked: when inserting a column left of B2, the note moved to C2; when inserting a row above C2, the not emoved to C3. However, my console gave some messages when I chose 'insert note' (via right mouse key - menu): warn:legacy.osl:21093:1:/home/winfried/git/libo/sfx2/source/control/shell.cxx:428: SfxShell::SetUndoManager: exchanging one non-NULL manager with another non-NULL manager? Suspicious! warn:legacy.osl:21093:1:/home/winfried/git/libo/drawinglayer/source/processor2d/vclhelperbufferdevice.cxx:64: impBufferDevice render helper: Copying from Window to VDev, this should be avoided (!) When ending inputting text into the note (selecting a cell at random), both messages appeared on my console again, this time in reversed order. I hope this helps. Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
RE: qa-build errros on clean code
Markus Mohrhard wrote (15 februari 2012 19:05) >Since you now disabled >the test can you at least check after a full build that if you start >calc insert a note in B2 and then insert a column before B2 the note >is moved to C2. Ideally you would even check that inserting a row will >move the note to C3. I can't check right now, but I will try later this afternoon. One thing that might help you: after performing git pull -r, the first build goes OK. When I change one little harmless thing in a source file (in sw in my case), make gives theassertion in ucalc.cxx. Reverting the change, make clean and make will still produce the assertion in ucalc.cxx. >And can you please tell us your build options. You are now the thirs >person in two weeks complaining about this failure so i would like to >fix it ASAP. My build option are: ./autogen.sh --with-max-jobs=6 --with-num-cpus=6 --without-junit --enable-dbgutil Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: qa-build errros on clean code
Hi Markus, Il 15/02/2012 19:05, Markus Mohrhard ha scritto: Hey, 2012/2/15 Riccardo Magliocchetti: Hi Winfried, Il 15/02/2012 16:30, Winfried Donkers ha scritto: Value: 8 33formel: =A1+B1 3 ucalc.cxx:2924:Assertion Test name: N12_GLOBAL__N_14TestE::testPostIts assertion failed - Expression: m_pDoc->GetNote(rAddr) == NULL - note hasn't moved Any suggestions? I have this very same failure plus another connectivity related, building with this patch applied to at least finish the build: http://paste.ubuntu.com/843064/ I don't like this solution but well the other option is to debug ucalc and find why it fails ( it does not for me ). Since you now disabled the test can you at least check after a full build that if you start calc insert a note in B2 and then insert a column before B2 the note is moved to C2. Ideally you would even check that inserting a row will move the note to C3. Yeah i know it is a bit lame from my side to disable tests and say "me too" on the list but i'm building a modified source which is crashing at runtime (see [RFC] x11 / cairo-less build try #3 mail [1]) and i am targeting headless so really it's not easy for me to debug the issue. And can you please tell us your build options. You are now the thirs person in two weeks complaining about this failure so i would like to fix it ASAP. This is x86-64 linux. ./autogen.sh --with-max-jobs=6 --with-num-cpus=3 --without-stlport --without-help --without-helppack-integration --without-myspell-dicts --without-junit --without-java --disable-epm --disable-gconf --disable-opengl --disable-gnome-vfs --disable-librsvg --disable-gtk --disable-gtk3 --disable-odk --disable-vba --disable-mozilla --disable-gstreamer --disable-randr --without-ppds --without-afms --disable-cairo-canvas --disable-neon --disable-dbus --disable-systray --disable-mathmldtd --disable-extension-integration --disable-postgresql-sdbc --with-system-libs --without-system-libcdr --without-system-libvisio --without-system-boost --without-system-libcmis --without-system-libwpd --without-system-cppunit --without-system-libwps --without-system-libwpg --disable-graphite --without-system-sampleicc --without-system-redland --without-system-hunspell --without-system-mythes --without-system-altlinuxhyph --without-system-lpsolve --without-system-libexttextcat --without-system-poppler --without-system-mdds --without-system-vigra --without-system-odbc --disable-ldap --disable-nsplugin --without-system-mozilla-headers --without-system-sane --disable-unix-qstart-libpng [1] http://lists.freedesktop.org/archives/libreoffice/2012-February/026170.html thanks -- Riccardo Magliocchetti ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: qa-build errros on clean code
Hey, 2012/2/15 Riccardo Magliocchetti : > Hi Winfried, > > Il 15/02/2012 16:30, Winfried Donkers ha scritto: >> >> Value: 8 >> 33formel: =A1+B1 >> >> 3 >> ucalc.cxx:2924:Assertion >> Test name: N12_GLOBAL__N_14TestE::testPostIts >> assertion failed >> - Expression: m_pDoc->GetNote(rAddr) == NULL >> - note hasn't moved >> Any suggestions? > > > I have this very same failure plus another connectivity related, building > with this patch applied to at least finish the build: > > http://paste.ubuntu.com/843064/ > I don't like this solution but well the other option is to debug ucalc and find why it fails ( it does not for me ). Since you now disabled the test can you at least check after a full build that if you start calc insert a note in B2 and then insert a column before B2 the note is moved to C2. Ideally you would even check that inserting a row will move the note to C3. And can you please tell us your build options. You are now the thirs person in two weeks complaining about this failure so i would like to fix it ASAP. Markus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
RE: qa-build errros on clean code
Riccardo Magliocchetti wrote (15 februari 2012) >> ucalc.cxx:2924:Assertion >I have this very same failure plus another connectivity related, >building with this patch applied to at least finish the build: >http://paste.ubuntu.com/843064/ Yes, I'll also disable the tests that produce assertions. Hopefully the cause will be found soon... Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: qa-build errros on clean code
Hi Winfried, Il 15/02/2012 16:30, Winfried Donkers ha scritto: Value: 8 33formel: =A1+B1 3 ucalc.cxx:2924:Assertion Test name: N12_GLOBAL__N_14TestE::testPostIts assertion failed - Expression: m_pDoc->GetNote(rAddr) == NULL - note hasn't moved Any suggestions? I have this very same failure plus another connectivity related, building with this patch applied to at least finish the build: http://paste.ubuntu.com/843064/ hth -- Riccardo Magliocchetti ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
RE: qa-build errros on clean code
>Tor already fixed it and after a pull you should be able to build. >(http://cgit.freedesktop.org/libreoffice/core/commit/?id=49a4b4c32bda86b97ef48d1d2ce9bf8f02d3e2e3) >Hope that your next build succeeds. >Markus No such luck :( After git pull -r, the first clean build went ok, but the next build gave the now familiar error: *** warn:legacy.osl:5863:1:/home/winfried/git/libo/sc/source/core/data/postit.cxx:640: ScPostIt::CreateCaptionFromInitData - need caption object or initial caption data Value: 8 33formel: =A1+B1 3 ucalc.cxx:2924:Assertion Test name: N12_GLOBAL__N_14TestE::testPostIts assertion failed - Expression: m_pDoc->GetNote(rAddr) == NULL - note hasn't moved Failures !!! Run: 30 Failure total: 1 Failures: 1 Errors: 0 warn:legacy.osl:5863:1:/home/winfried/git/libo/unotools/source/config/configmgr.cxx:217: OSL_ASSERT 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[1]: *** [/home/winfried/git/libo/workdir/unxlngi6/CppunitTest/sc_ucalc.test] Error 1 make[1]: *** Waiting for unfinished jobs dmake: Error code 2, while making 'all' *** Any suggestions? Winfried ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: qa-build errros on clean code
Hey, > I am experiencing problems with building (openSUSE 12.1). > > I have two machine that can build the code. > > One of the machines started acting up last weekend with assertions from > sc/qa/unit/ucalc.cxx when performings tests. Removing all my changes and > updating the code (./g pull -r) does not help. I used autogen.sh with > parameters with-max-jobs=6 --with-num-cpus=6 --without-junit > --enable-dbgutil (and sometimes --enable-debug). > > I even produced a brand new clone with git; same error. > > > > The other produced an error after ./g pull -r yesterday. On this one I used > autogen.sh with parameters with-max-jobs=1 --with-num-cpus=1 --without-junit > --enable-dbgutil. > > The error message is: > > /home/w.donkers/git/libo/sw/qa/extras/rtftok/rtftok.cxx: In member function > 'void RtfModelTest::testFdo45543()': > > /home/w.donkers/git/libo/sw/qa/extras/rtftok/rtftok.cxx:159:5: error: no > matching function for call to 'assertEquals(int, sal_Int32, > CppUnit::SourceLine, const char [1])' > Tor already fixed it and after a pull you should be able to build. (http://cgit.freedesktop.org/libreoffice/core/commit/?id=49a4b4c32bda86b97ef48d1d2ce9bf8f02d3e2e3) Hope that your next build succeeds. Markus ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice