On 10/07/12 16:42, Miklos Vajna wrote: > Hi, > > I try to run make check before pushing from time to time (and when it > fails, I usually refer to Stephan's tinderbox to see if I can ignore the > failure), but since commit c569add9b9b9e5aadff4b1379d1848c192f027ee, > make check now fails, even if the mentioned tinderbox has no problem.
> Thread 1 (Thread 0x2ac447a45700 (LWP 16584)): > #0 0x00002ac43b556d95 in raise () from /lib64/libc.so.6 > #1 0x00002ac43b5582ab in abort () from /lib64/libc.so.6 > #2 0x00002ac43b598700 in malloc_printerr () from /lib64/libc.so.6 malloc? perhaps some memory management problem? i'd suggest running the test under valgrind, like so: (cd chart2 && make $(readlink -f ../workdir/unxlngx6/JunitTest/chart2_unoapi/done) VALGRIND=memcheck) this should now work on master with fb79f6d64eee0580558dac15f42f97ee2255ff24, where i've removed the useless --error-exitcode that prevented the test from actually running with valgrind. > I tried to check what can be the problem, it turns out that: > > LANG=hu_HU.UTF-8 LC_ALL= make -sr > /home/vmiklos/git/libreoffice/master/workdir/unxlngx6.pro/JunitTest/chart2_unoapi/done > > fails, but > > LANG= LC_ALL= make -sr > /home/vmiklos/git/libreoffice/master/workdir/unxlngx6.pro/JunitTest/chart2_unoapi/done > > passes fine. Installed dictionaries (from openSUSE 12.1 packages): > > $ rpm -qa|grep myspell- > myspell-american-20100316-24.1.2.noarch > myspell-hungarian-20100329-24.1.2.noarch > > As far as I see, these system dictionaries are picked up by default by > configure. I guess, the real problem can be in hnj_hyphen_hyph_(), so > maybe ff3c4e9e54efff3b8701fb562291a9edacae8100 (hyphen upgrade) is to be > blamed? hmmm probably not many people run the unit tests in a hungarian locale :) _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
