Hi Enrico, [also Cc'ing Fabian and the relevant bug report]
On Thu, Feb 18, 2010 at 02:26:27PM +0100, Enrico Zini wrote:
> On Thu, Feb 18, 2010 at 12:49:21PM +0100, Philipp Kern wrote:
> > would you object or welcome a direct NMU for #570306? I just want guessnet
> > installable in unstable again and the binNMUs failed badly. ;-)
> > That still leaves us with #553906 which could've been transient. \-:
> Thanks! I have no problem at all with a NMU. In fact, there's even an
> open RFA for guessnet, because since I'm not using it anymore I have a
> bit of a hard time to test a new build before uploading.
I just made a build on sparc (sadly the porterbox is down and all that's left
for me to test on is a buildd), and there seems to be a race in the testsuite.
The first build failed in the testsuite. Sadly I didn't preverse the build
tree and when I tried that it didn't fail anymore. The failure is the same as
in #553906, but the build terminates normally:
>>> snip >>>
scanner_scanbag: ....
starter: ...
Segmentation fault
util_processrunner: .FAIL: tests/guessnet-test
==================================
1 of 1 test failed
Please report to enr...@debian.org
>>> snip >>>
What I get with `make check' at the toplevel is this:
>>> snip >>>
scanner_scanbag: ....
starter: ...
util_processrunner: [1=F]
---> group: util_processrunner, test: test<1>
problem: assertion failed
failed assertion: "util/processrunner-tut.cc:86(lfalse.tag == "false"): :
expected 'false' actual ''"
tests summary: failures:1 ok:7
PASS: tests/guessnet-test
>>> snip >>>
The next time I try I get this:
>>> snip >>>
scanner_scanbag: ....
starter: ...
util_processrunner: [1=F]killing signal catched
Bus error
FAIL: tests/guessnet-test
>>> snip >>>
gdb yields this sometimes (it's definitely a race):
>>> snip >>>
util_processrunner: .
Program received signal SIGBUS, Bus error.
[Switching to Thread 0xf7ad7b70 (LWP 13886)]
0xf7e08424 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string(std::string const&) () from
/usr/lib/libstdc++.so.6
(gdb) bt
#0 0xf7e08424 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string(std::string const&) () from
/usr/lib/libstdc++.so.6
#1 0x00046e94 in ProcData (this=0xbb8e0) at util/processrunner.h:48
#2 ProcessRunner::main (this=0xbb8e0) at util/processrunner.cc:215
#3 0x0007dd44 in wibble::sys::Thread::Starter(void*) ()
#4 0xf7ec6358 in start_thread () from /lib/libpthread.so.0
#5 0xf7bc5f1c in ?? () from /lib/libc.so.6
#6 0xf7bc5f1c in ?? () from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
>>> snip >>>
I cannot really step through the code because of the threading.
The relevant code snippet (if we trust the backtrace):
>>> snip >>>
212 while (!queuedForRunning.empty())
213 {
214 debug("PRI Run a queued
process\n");
215 ProcData d =
queuedForRunning.front();
216 queuedForRunning.pop();
217 d.run();
218
proclist.insert(make_pair(d.script->pid(), d));
219 //debug("run process %s
pid: %d\n", d.tag.c_str(), d.script->pid());
220 }
>>> snip >>>
Other failures:
>>> snip >>>
scanner_scanbag: ....
starter: ...
util_processrunner: [1=F]terminate called after throwing an instance of
'std::runtime_error'
what(): killing signal catched
Aborted (core dumped)
>>> snip >>>
That gives us a garbled stack. All the core dumps I'm able to generate
are not useful.
One backtrace I'm able to get in gdb directly:
>>> snip >>>
(gdb) bt
#0 0xf7e6d070 in ?? () from /usr/lib/libstdc++.so.6
#1 0xf7c997bc in _Unwind_DeleteException () from /lib/libgcc_s.so.1
#2 0x0003d8a0 in tut::test_runner::run_tests (argc=<value optimized out>,
argv=<value optimized out>) at /usr/include/wibble/tests/tut.h:318
#3 main (argc=<value optimized out>, argv=<value optimized out>) at
tests/tut-main.cpp:43
>>> snip >>>
Let's go `-O0 -g':
>>> snip >>>
util_processrunner: .
Program received signal SIGBUS, Bus error.
[Switching to Thread 0xf7b07b70 (LWP 29473)]
0xf7e38424 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string(std::string const&) () from
/usr/lib/libstdc++.so.6
(gdb) bt
#0 0xf7e38424 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string(std::string const&) () from
/usr/lib/libstdc++.so.6
#1 0x0007362c in ProcData (this=0xf7b071dc) at util/processrunner.h:48
#2 0x000704fc in ProcessRunner::main (this=0x11b8e0) at
util/processrunner.cc:215
#3 0x000bc79c in wibble::sys::Thread::Starter(void*) ()
#4 0xf7ef6358 in start_thread () from /lib/libpthread.so.0
#5 0xf7bf5f1c in ?? () from /lib/libc.so.6
#6 0xf7bf5f1c in ?? () from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb) up 1
#1 0x0007362c in ProcData (this=0xf7b071dc) at util/processrunner.h:48
48 {
(gdb) bt full
#0 0xf7e38424 in std::basic_string<char, std::char_traits<char>,
std::allocator<char> >::basic_string(std::string const&) () from
/usr/lib/libstdc++.so.6
No symbol table info available.
#1 0x0007362c in ProcData (this=0xf7b071dc) at util/processrunner.h:48
No locals.
#2 0x000704fc in ProcessRunner::main (this=0x11b8e0) at
util/processrunner.cc:215
d = {tag = {static npos = 4294967295,
_M_dataplus = {<std::allocator<char>> =
{<__gnu_cxx::new_allocator<char>> = {<No data fields>}, <No data fields>}, _M_p
= 0x11b998 ""}},
script = 0x11c7e8, listener = 0x11b610}
lock = {mutex = @0x11b8ec, locked = true, yield = false}
pid = 0
status = 256
want_shutdown = false
sigs = {__val = {2147479879, 4294967294, 4294967295 <repeats 30 times>}}
oldsigs = {__val = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2020, 976,
4026531840, 4155535624, 4155555844, 4155550328, 0, 0, 5, 1093, 4160297568,
4160295888,
4159644435, 4155595396, 4159639492, 1, 0, 2822930839, 4160293000,
4155535852, 4155535840, 1}}
proclist = {_M_t = {
_M_impl = {<std::allocator<std::_Rb_tree_node<std::pair<int const,
processrunner::ProcData> > >> =
{<__gnu_cxx::new_allocator<std::_Rb_tree_node<std::pair<int const,
processrunner::ProcData> > >> = {<No data fields>}, <No data fields>},
_M_key_compare = {<std::binary_function<int, int, bool>> = {<No
data fields>}, <No data fields>}, _M_header = {_M_color = std::_S_red,
_M_parent = 0x11b5a8, _M_left = 0x11b5a8, _M_right = 0x11bdd8},
_M_node_count = 2}}}
#3 0x000bc79c in wibble::sys::Thread::Starter(void*) ()
No symbol table info available.
#4 0xf7ef6358 in start_thread () from /lib/libpthread.so.0
No symbol table info available.
#5 0xf7bf5f1c in ?? () from /lib/libc.so.6
No symbol table info available.
#6 0xf7bf5f1c in ?? () from /lib/libc.so.6
No symbol table info available.
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
>>> snip >>>
Sadly that doesn't look more helpful and I'm now utterly confused how to
continue. :)
Kind regards,
Philipp Kern
--
.''`. Philipp Kern Debian Developer
: :' : http://philkern.de Stable Release Manager
`. `' xmpp:p...@0x539.de Wanna-Build Admin
`- finger pkern/k...@db.debian.org
signature.asc
Description: Digital signature
