On Mon, Dec 12, 2016 at 09:54:57PM +0100, Emilio Pozuelo Monfort wrote: > Huh, this is crashing in libc's malloc(). > > Can you run the test under valgrind's memcheck? Maybe there's some memory > corruption that would explain this.
Do you mean this? valgrind the-test-that-we-want-to-debug How do I run only the test that we want to debug? (I got the previus backtrace by pure chance, after "ulimit -c unlimited" and "make check"). Do I need any LD_LIBRARY_PATH or something? Any command line arguments? If I just do this: valgrind tests/data-models/.libs/check_data_proxy I get this: ==25068== Memcheck, a memory error detector ==25068== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al. ==25068== Using Valgrind-3.12.0 and LibVEX; rerun with -h for copyright info ==25068== Command: tests/data-models/.libs/check_data_proxy ==25068== tests/data-models/.libs/check_data_proxy: error while loading shared libraries: libgda-5.0.so.4: cannot open shared object file: No such file or directory ==25068== Jump to the invalid address stated on the next line ==25068== at 0x596: ??? ==25068== by 0x400F352: _dl_signal_error (dl-error.c:125) ==25068== by 0x400DF05: _dl_map_object_deps (dl-deps.c:686) ==25068== by 0x4002EF8: dl_main (rtld.c:1610) ==25068== by 0x401766D: _dl_sysdep_start (dl-sysdep.c:249) ==25068== by 0x4004D27: _dl_start_final (rtld.c:307) ==25068== by 0x4004D27: _dl_start (rtld.c:413) ==25068== by 0x4000CA7: ??? (in /lib/x86_64-linux-gnu/ld-2.24.so) ==25068== Address 0x596 is not stack'd, malloc'd or (recently) free'd ==25068== ==25068== ==25068== Process terminating with default action of signal 11 (SIGSEGV): dumping core ==25068== Bad permissions for mapped region at address 0x596 ==25068== at 0x596: ??? ==25068== by 0x400F352: _dl_signal_error (dl-error.c:125) ==25068== by 0x400DF05: _dl_map_object_deps (dl-deps.c:686) ==25068== by 0x4002EF8: dl_main (rtld.c:1610) ==25068== by 0x401766D: _dl_sysdep_start (dl-sysdep.c:249) ==25068== by 0x4004D27: _dl_start_final (rtld.c:307) ==25068== by 0x4004D27: _dl_start (rtld.c:413) ==25068== by 0x4000CA7: ??? (in /lib/x86_64-linux-gnu/ld-2.24.so) ==25068== ==25068== HEAP SUMMARY: ==25068== in use at exit: 0 bytes in 0 blocks ==25068== total heap usage: 0 allocs, 0 frees, 0 bytes allocated ==25068== ==25068== All heap blocks were freed -- no leaks are possible ==25068== ==25068== For counts of detected and suppressed errors, rerun with: -v ==25068== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0) and because of the "error while loading shared libraries" part I'm not sure this is exactly what should be done. Well, I finally figured out (maybe). export LD_LIBRARY_PATH=pwd/libgda/.libs cd tests/data-models valgrind .libs/check_data_proxy and this is the valgrind report at the end: ==25768== ==25768== HEAP SUMMARY: ==25768== in use at exit: 1,730,127 bytes in 45,645 blocks ==25768== total heap usage: 809,031 allocs, 763,386 frees, 736,853,843 bytes allocated ==25768== ==25768== LEAK SUMMARY: ==25768== definitely lost: 31,346 bytes in 253 blocks ==25768== indirectly lost: 7,898 bytes in 241 blocks ==25768== possibly lost: 2,638 bytes in 27 blocks ==25768== still reachable: 1,558,413 bytes in 43,683 blocks ==25768== of which reachable via heuristic: ==25768== length64 : 440 bytes in 11 blocks ==25768== newarray : 1,568 bytes in 18 blocks ==25768== suppressed: 0 bytes in 0 blocks ==25768== Rerun with --leak-check=full to see details of leaked memory ==25768== ==25768== For counts of detected and suppressed errors, rerun with: -v ==25768== ERROR SUMMARY: 40 errors from 3 contexts (suppressed: 0 from 0) Is this what you need? Thanks.
