On Sat, Jul 2, 2011 at 01:03:05 +0200, Julien Cristau wrote: > On Fri, Jul 1, 2011 at 14:08:08 +0100, Simon McVittie wrote: > > > Package: telepathy-glib > > Version: 0.15.1-1 > > Severity: serious > > Justification: fails to build from source > > > > The feature test test-connection-balance reliably segfaults on ia64. I don't > > think we've seen this failure on any other architecture. It might be to do > > with word size, a race condition, or some sort of undefined behaviour that > > happens to give us the right result on every other architecture. > > > Reproduced on merulo, got a core file, unfortunately no glib debug > symbols yet so this isn't as useful as it could be: > Fixed now:
(gdb) bt full
#0 0x20000000007984c1 in g_type_fundamental (type_id=<value optimized out>)
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./gobject/gtype.c:3946
node = 0x1ec20
#1 0x2000000000764cf0 in g_object_new_valist (object_type=125984,
first_property_name=0x400000000000e538 "account",
var_args=0x60000fffffd2ac30)
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./gobject/gobject.c:1552
params = <value optimized out>
name = <value optimized out>
object = 0x0
n_params = <value optimized out>
n_alloced_params = <value optimized out>
__PRETTY_FUNCTION__ = "g_object_new_valist"
#2 0x400000000000d000 in tp_tests_object_new_static_class (type=125984)
at util.c:267
No locals.
#3 0x4000000000007d10 in setup (test=0x60000000000203e0,
data=0x600000000001ec20) at connection-balance.c:245
error = 0x3b
features = {59, 0}
__PRETTY_FUNCTION__ = "setup"
#4 0x2000000000909b00 in test_case_run (suite=0x6000000000005ec0,
path=0x20000000009a81c0 "")
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./glib/gtestutils.c:1173
test_run_timer = 0x60000000000203a0
largs = {<invalid float value>, 0, -inf}
fixture = 0x60000000000203e0
old_name = 0x6000000000020340 "/conn"
old_base = 0x0
#5 g_test_run_suite_internal (suite=0x6000000000005ec0,
path=0x20000000009a81c0 "")
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./glib/gtestutils.c:1223
tc = 0x600000000001b830
bad_suite = <value optimized out>
l = 0
rest = 0x0
old_name = 0x6000000000020290 ""
slist = 0x6000000000011170
reversed = 0x6000000000011170
__PRETTY_FUNCTION__ = "g_test_run_suite_internal"
#6 0x2000000000909860 in g_test_run_suite_internal (
suite=<value optimized out>, path=0x20000000009a81c0 "")
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./glib/gtestutils.c:1233
ts = 0x6000000000005ec0
bad_suite = 0
l = 0
rest = 0x20000000009a81c0 ""
old_name = 0x20000000009a81c0 ""
slist = <value optimized out>
reversed = 0x6000000000011160
__PRETTY_FUNCTION__ = "g_test_run_suite_internal"
#7 0x2000000000909ec0 in g_test_run_suite (suite=0x6000000000005ea0)
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./glib/gtestutils.c:1282
rest = 0x6000000000011130 "\260\020\001"
path = 0x20000000009a81c0 ""
l = <value optimized out>
n = 0
n_bad = 0
__PRETTY_FUNCTION__ = "g_test_run_suite"
#8 0x2000000000909fb0 in g_test_run ()
at
/build/buildd-glib2.0_2.28.6-1-ia64-jftWAX/glib2.0-2.28.6/./glib/gtestutils.c:878
No locals.
#9 0x4000000000008560 in main (argc=2, argv=0x2) at connection-balance.c:427
No locals.
One suspicious thing is that the test's setup function takes a pointer
(data=0x600000000001ec20), and converts it to uint (conn_type =
GPOINTER_TO_UINT(data) = 0x1ec20 = 125984), truncating it?
Cheers,
Julien
signature.asc
Description: Digital signature
