Segfault with postgres provider
Submitted by Piotr Pokora
Assigned to mal..@..db.org
Link to original bug (#682459)
Description
Midgard uses travis fot tests. The same tests fail with postgres provider: http://travis-ci.org/#!/piotras/midgard-core/builds/2198990
Unfortunately I am not able to provide standalone use case.
So far, I managed to get this backtrace:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xae3ffb40 (LWP 20444)]
0xb6cbe13a in resetPQExpBuffer () from /usr/lib/libpq.so.5
(gdb) bt
#0 0xb6cbe13a in resetPQExpBuffer () from /usr/lib/libpq.so.5
#1 0xb6cbe252 in printfPQExpBuffer () from /usr/lib/libpq.so.5
#2 0xb6cb46aa in PQgetResult () from /usr/lib/libpq.so.5
#3 0xb6cb4877 in ?? () from /usr/lib/libpq.so.5
#4 0xb6cb4cf9 in PQexec () from /usr/lib/libpq.so.5
#5 0xb6ce34d4 in _gda_postgres_PQexec_wrap (cnc=0x8b87390, pconn=0x8de5be0, query=0x8b32068 "DEALLOCATE psc561")
at gda-postgres-util.c:141
#6 0xb6ce11ad in gda_postgres_pstmt_finalize (object=0x9159c08) at gda-postgres-pstmt.c:95
#7 0xb7676288 in g_object_unref (_object=0x9159c08) at /build/buildd/glib2.0-2.32.1/./gobject/gobject.c:3018
#8 0xb7052daa in gda_data_select_dispose (object=0xae60a950) at gda-data-select.c:489
#9 0xb6ce2e19 in gda_postgres_recordset_dispose (object=0xae60a950) at gda-postgres-recordset.c:187
#10 0xb76761a2 in g_object_unref (_object=0xae60a950) at /build/buildd/glib2.0-2.32.1/./gobject/gobject.c:2981
#11 0xb70dfed7 in worker_thread_entry_point (to_worker_thread=0x8e0bfd0) at gda-thread-wrapper.c:620
#12 0xb75d5673 in g_thread_proxy (data=0x8bbe6c0) at /build/buildd/glib2.0-2.32.1/./glib/gthread.c:801
#13 0xb7fa3d4c in start_thread (arg=0xae3ffb40) at pthread_create.c:308
#14 0xb7cc1ace in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:130
Libgda is 5.1.1, but the same segfault happens with 4.x.
I assume this is postgres provider bug, as it never happens with SQLite or MySQL.
Version: 5.1.x