Re: 3.3-rc1 crash bug (?)
At 1242331287 time_t, Uli Schlachter wrote: And another backtrace, this time with gdb... The plan for is to finally use -O0 -fno-inline and see what happens. (btw yay, finally done with school and I got time for this!) (Oh and another btw: I still don't know any way to reproduce this, it just happens :( ) Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fe166b55740 (LWP 4868)] 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 #0 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 No symbol table info available. #1 0x0032a5c7b92c in memmove () from /lib/libc.so.6 No symbol table info available. #2 0x00439e25 in luaA_imagebox_newindex (L=0x26e10b0, token=value optimized out) at /home/ancient/jd/Work/debian/awesome/common/luaobject.h:39 buf = value optimized out len = value optimized out widget = (widget_t *) 0x2b3bcd8 d = (imagebox_data_t *) 0x2937550 #3 0x00430032 in luaA_widget_newindex (L=0x26e10b0) at /home/ancient/jd/Work/debian/awesome/widget.c:490 len = 5 widget = (widget_t *) 0x2b3bcd8 buf = value optimized out token = A_TK_UNKNOWN That's just weird. If the token is A_TK_UNKNOWN, it's because some code is doing myimagebox.blabla = value where blabla is not a known attribute, therefore we get A_TK_UNKNOWN as token. Then it calls luaA_imagebox_newindex which does not handle A_TK_UNKNOWN in its switch() statement, so it goes to default which return 0. Where the hell memmove is called from lua object function, I don't see. I suggest that next time you try to print some stuff and dig directly into the code/debugger because I really miss what's wrong. Thanks anyway :) Cheers, -- Julien Danjou // ᐰ jul...@danjou.info http://julien.danjou.info // 9A0D 5FD9 EB42 22F6 8974 C95C A462 B51E C2FE E5CD // Tomorrow I was nothing, yesterday I'll be. signature.asc Description: Digital signature
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Julien Danjou wrote: At 1242331287 time_t, Uli Schlachter wrote: And another backtrace, this time with gdb... The plan for is to finally use -O0 -fno-inline and see what happens. (btw yay, finally done with school and I got time for this!) (Oh and another btw: I still don't know any way to reproduce this, it just happens :( ) Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fe166b55740 (LWP 4868)] 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 #0 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 No symbol table info available. #1 0x0032a5c7b92c in memmove () from /lib/libc.so.6 No symbol table info available. #2 0x00439e25 in luaA_imagebox_newindex (L=0x26e10b0, token=value optimized out) at /home/ancient/jd/Work/debian/awesome/common/luaobject.h:39 buf = value optimized out len = value optimized out widget = (widget_t *) 0x2b3bcd8 d = (imagebox_data_t *) 0x2937550 #3 0x00430032 in luaA_widget_newindex (L=0x26e10b0) at /home/ancient/jd/Work/debian/awesome/widget.c:490 len = 5 widget = (widget_t *) 0x2b3bcd8 buf = value optimized out token = A_TK_UNKNOWN That's just weird. If the token is A_TK_UNKNOWN, it's because some code is doing myimagebox.blabla = value where blabla is not a known attribute, therefore we get A_TK_UNKNOWN as token. This call is from somewhere inside the tasklist widget, so I'd expect no such weirdness in there. Then it calls luaA_imagebox_newindex which does not handle A_TK_UNKNOWN in its switch() statement, so it goes to default which return 0. Where the hell memmove is called from lua object function, I don't see. I suggest that next time you try to print some stuff and dig directly into the code/debugger because I really miss what's wrong. Thanks anyway :) Cheers, Well, I still don't have a test case for this. It just... happens... sometimes. Oh and since debian testing now got kde4 I don't have kde3's konqueror anymore. I bet this fixes this bug. Meh... :( Cheers, Uli P.S.: Welcome back, jd. - -- Do you know that books smell like nutmeg or some spice from a foreign land? -- Faber in Fahrenheit 451 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEcBAEBCAAGBQJKEVuJAAoJECLkKOvLj8sGDO4H/RPiTvLQ4FsPArr3We5FLjcW jvZSBwxqY+mOgL0KPipGU/0gIRylUdMkrs+0c1m8Q5kh0a5MuWKKB27uqW6IOQyd lug73YMNVetA7afmXpnnmqABt0ekSAnA+ESiJh/JffWI20iu4kV5qy9mz7ser6bh UK4lycgGpayZh6HrE2sh2I+7Q7E1olAWi+sGQqFSbHSeb0pXYm62D3DPZEF2srUZ 4+B5ts1+7EdwgsR2Mla38jIalSJBPSLwvzBv77lYrxuDkvl3QpJDuH8th2VZ8jO7 PgiVnc4/GqcPIICUZi32K2S3/UpM+kTO53evmABbrZooPjUnOEwHGI2fBPieDsA= =QLIi -END PGP SIGNATURE- -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Nikos Ntarmos wrote: Hi there. FWIW I've also been experiencing segfaults on FreeBSD/i386, however only when X goes down before awesome. A short look into this revealed it was due to widgets update hooks being scheduled for some time after awesome and X is down. I never got around to looking into this any further but it's still there in my todo list. If you disable any widgets, timers, etc. do you still get those backtraces? Cheers. \n\n Hi, Sorry that I didn't reply earlier, I kinda missed this. This sounds like a completely different issue to me, since my crashes are during normal usage, nor during shutdown. Cheers, Uli - -- Do you know that books smell like nutmeg or some spice from a foreign land? -- Faber in Fahrenheit 451 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEcBAEBCAAGBQJKDHKPAAoJECLkKOvLj8sG83sH/jZ+mabrJkydEJumflbT72tc 9VdgnRz3L1ZmbgG5IFn7+q10Rm7QE9njxstkJJFziB0JnGO8Ll8KBmkUu5I5N8fI DMb6Xj4amyCFr86OwVgQMd6fPlGHWWoyd+i3H274QWR8ARUrjiQXKDSlcm/I4xo/ isuqiwrj2soKQtTYGuvKL/uxRw/m+X6kKx86Xif8wpBg2tZkug6DfQQE8Z+XbUaN EyGNDvL2WwRtNX0nD5seCuYimGYzezNndwwrMkTKnQyNj+7tLkuZEpk68oTZtqQs +nt/0OV9g0xNVcFLgwpvCkPp8uDCdpiRbXNj1tzs/2hdp22X2KN3tFwh/LSJakg= =A2xw -END PGP SIGNATURE- -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Uli Schlachter wrote: Uli Schlachter wrote: Julien Danjou wrote: At 1241874772 time_t, Uli Schlachter wrote: Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( http://naquadah.org/~jd/debian/awesome_3.3~rc2-1_nostrip-debug-noopt_amd64.deb Hi, now it crashed again and valgrind finally gave some more useful backtraces: ==4386== Invalid read of size 1 ==4386==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4386==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4386== Address 0x5cde9b0 is 0 bytes after a block of size 96 alloc'd ==4386==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4386==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4386==by 0x412EA2: client_manage (util.h:155) ==4386==by 0x417D90: event_handle_maprequest (event.c:645) ==4386==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4386==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4386==by 0x40E8A6: main (awesome.c:536) New day, new backtrace (why does it only ever crash once per day for me?). Notice that the memory was allocated in 'property_handle_net_wm_icon (util.h:155)'. util.h:155 ist xrealloc() which is only ever called from p_realloc() which in turn is called from several places. Looking through client_manage() I would expect either the call to client_array_push() or the image_ref() call (which calls luaA_ref_array_append()) to be the cause. So this all points to the code in common/array.h, I guess, but I'd expect bugs in there to be more visible... :/ ==4485== Invalid read of size 1 ==4485==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==4485==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4485==by 0x430031: luaA_widget_newindex (widget.c:490) ==4485==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4485==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4485== Address 0x8136948 is 0 bytes after a block of size 96 alloc'd ==4485==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4485==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4485==by 0x41A923: property_handle_net_wm_icon (util.h:155) ==4485==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4485==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4485==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4485==by 0x40E8A6: main (awesome.c:536) Cheers, Uli P.S. I'm only posting this so that I can find all my backtraces in one place. And another backtrace, this time with gdb... The plan for is to finally use -O0 -fno-inline and see what happens. (btw yay, finally done with school and I got time for this!) (Oh and another btw: I still don't know any way to reproduce this, it just happens :( ) Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fe166b55740 (LWP 4868)] 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 #0 0x0032a5c7d4dc in ?? () from /lib/libc.so.6 No symbol table info available. #1 0x0032a5c7b92c in memmove () from /lib/libc.so.6 No symbol table info available. #2 0x00439e25 in luaA_imagebox_newindex (L=0x26e10b0, token=value optimized out) at /home/ancient/jd/Work/debian/awesome/common/luaobject.h:39 buf = value optimized out len = value optimized out widget = (widget_t *) 0x2b3bcd8 d = (imagebox_data_t *) 0x2937550 #3 0x00430032 in luaA_widget_newindex (L=0x26e10b0) at /home/ancient/jd/Work/debian/awesome/widget.c:490 len = 5 widget = (widget_t *) 0x2b3bcd8 buf = value optimized out
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Uli Schlachter wrote: Julien Danjou wrote: At 1241874772 time_t, Uli Schlachter wrote: Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( http://naquadah.org/~jd/debian/awesome_3.3~rc2-1_nostrip-debug-noopt_amd64.deb Hi, now it crashed again and valgrind finally gave some more useful backtraces: ==4386== Invalid read of size 1 ==4386==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4386==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4386== Address 0x5cde9b0 is 0 bytes after a block of size 96 alloc'd ==4386==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4386==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4386==by 0x412EA2: client_manage (util.h:155) ==4386==by 0x417D90: event_handle_maprequest (event.c:645) ==4386==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4386==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4386==by 0x40E8A6: main (awesome.c:536) New day, new backtrace (why does it only ever crash once per day for me?). Notice that the memory was allocated in 'property_handle_net_wm_icon (util.h:155)'. util.h:155 ist xrealloc() which is only ever called from p_realloc() which in turn is called from several places. Looking through client_manage() I would expect either the call to client_array_push() or the image_ref() call (which calls luaA_ref_array_append()) to be the cause. So this all points to the code in common/array.h, I guess, but I'd expect bugs in there to be more visible... :/ ==4485== Invalid read of size 1 ==4485==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==4485==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4485==by 0x430031: luaA_widget_newindex (widget.c:490) ==4485==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4485==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4485==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4485== Address 0x8136948 is 0 bytes after a block of size 96 alloc'd ==4485==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4485==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4485==by 0x41A923: property_handle_net_wm_icon (util.h:155) ==4485==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4485==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4485==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4485==by 0x40E8A6: main (awesome.c:536) Cheers, Uli P.S. I'm only posting this so that I can find all my backtraces in one place. - -- Do you know that books smell like nutmeg or some spice from a foreign land? -- Faber in Fahrenheit 451 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEcBAEBCAAGBQJKCR3/AAoJECLkKOvLj8sGvEAH/jPd7aPZFCXAfk55cC7i/DaL i0F5SBz5ls5E+AUPxwCm04SN7XSD0jsUSf9X2Hop22d33+psrhJitayKsapj8lxT wzZEn3rpNTMbxu1IWOJyWWaGN+XVyJXIdOGGssTYN5Fus070t2tvX58ergDfi6h2 kuXnrfshoUxkOhu3XLi3HQafJby8SDQxmh/9J/7GqOB2n73oyorZMBgWYOBolUBR RziJpqo6XvXgAkOa3H+ufMwCReCGmaEY/nADelkcYUP4C+P/nIuaXjhN/fUEa3FM pvj4hX6zy3hctv1s01o+gQ6YGc3WEIMso5jLhBPx5JTb87iYTFWRT6t4TEmSLWw= =p9LU -END PGP SIGNATURE- -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
Hi there. FWIW I've also been experiencing segfaults on FreeBSD/i386, however only when X goes down before awesome. A short look into this revealed it was due to widgets update hooks being scheduled for some time after awesome and X is down. I never got around to looking into this any further but it's still there in my todo list. If you disable any widgets, timers, etc. do you still get those backtraces? Cheers. \n\n -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Julien Danjou wrote: At 1241874772 time_t, Uli Schlachter wrote: Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( http://naquadah.org/~jd/debian/awesome_3.3~rc2-1_nostrip-debug-noopt_amd64.deb Hi, now it crashed again and valgrind finally gave some more useful backtraces: ==4386== Invalid read of size 1 ==4386==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4386==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4386== Address 0x5cde9b0 is 0 bytes after a block of size 96 alloc'd ==4386==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4386==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4386==by 0x412EA2: client_manage (util.h:155) ==4386==by 0x417D90: event_handle_maprequest (event.c:645) ==4386==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4386==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4386==by 0x40E8A6: main (awesome.c:536) ==4386== ==4386== Invalid read of size 1 ==4386==at 0x4A09F90: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4386==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4386== Address 0x5cde9b1 is 1 bytes after a block of size 96 alloc'd ==4386==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4386==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4386==by 0x412EA2: client_manage (util.h:155) ==4386==by 0x417D90: event_handle_maprequest (event.c:645) ==4386==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4386==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4386==by 0x40E8A6: main (awesome.c:536) ==4386== ==4386== Invalid write of size 1 ==4386==at 0x4A09F94: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x41A843: property_handle_net_wm_icon (luaa.h:261) ==4386==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==4386== Address 0x5cde9b0 is 0 bytes after a block of size 96 alloc'd ==4386==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==4386==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==4386==by 0x412EA2: client_manage (util.h:155) ==4386==by 0x417D90: event_handle_maprequest (event.c:645) ==4386==by 0x40DFDC: a_xcb_check_cb (awesome.c:203) ==4386==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==4386==by 0x40E8A6: main (awesome.c:536) ==4386== ==4386== Process terminating with default action of signal 11 (SIGSEGV) ==4386== Bad permissions for mapped region at address 0x605B000 ==4386==at 0x4A09F94: memmove (mc_replace_strmem.c:517) ==4386==by 0x439E24: luaA_imagebox_newindex (luaobject.h:39) ==4386==by 0x430031: luaA_widget_newindex (widget.c:490) ==4386==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==4386==by
3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hi, no idea what exactly just happened, but awesome crashed when I pressed ctrl-q in konqueror. Syslog says: May 9 10:49:21 psytux kernel: [ 8809.790642] awesome[4487]: segfault at 257a000 ip 0032a5c7d511 sp 7fffbd55e868 error 4 in libc-2.9.so[32a5c0+149000] (Any idea how I can find out what error 4 means? Uhm, ok, so where is this? objdump -S /lib/libc-2.9.so | vim -R - And I let vim search for c7d511 (the last part of the instruction pointer shown): 32a5c7d511: 4d 8b 44 24 10 mov0x10(%r12),%r8 Scrolling up a little revealed that the next visible symbol there is memcpy(): 0032a5c7ced0 memcpy: I have no idea what caused this and it doesn't seem reproducible which is why I won't open a bug report. I'm updating to rc2 now and I will run awesome in valgrind. Let's see if I can find anything... Cheers, Uli - -- Do you know that books smell like nutmeg or some spice from a foreign land? -- Faber in Fahrenheit 451 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEcBAEBCAAGBQJKBU3EAAoJECLkKOvLj8sGYYcH/3k6qnEY/TBJIMieqJ6/BloW 3NJgvC2kmcsDuPD6ETJt2ywA/xcCm/p8Li2lYUd1STs3DaML2vrL14HZdFhdU0DR i5gz6HNzWXWx9d1a8544moWg3nsfRbnuRhV75FQtx2U8NN7ZSKgj20ElHKifKItf 2p61xMSNKI0/jlg0NZqcVwUPe6+eexsS1Ouqj5UBftxeTdCFpcNIn0BX0AcW6K9X Cds+QkzcMEhNAZXMM4yafhti94cSXxAKv+4wNAgPTQ5QA+eMYVegxkh8Od+KqAwJ xQpEFwUvq39L0hJd3flrZ199soIKQGQMl5iIRMs7MiOMSZfq1YPsLhIeBqI7F+0= =MbhN -END PGP SIGNATURE- -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
At 1241861577 time_t, Uli Schlachter wrote: I have no idea what caused this and it doesn't seem reproducible which is why I won't open a bug report. I'm updating to rc2 now and I will run awesome in valgrind. Let's see if I can find anything... What I can suggest is to call ulimit -c unlimited before running awesome. At least you'll have a coredump. And if you can reproduce it, recompiling with -O0 -fno-inline can help to have a more complete bt. Cheers, -- Julien Danjou // ᐰ jul...@danjou.info http://julien.danjou.info // 9A0D 5FD9 EB42 22F6 8974 C95C A462 B51E C2FE E5CD // Trust me. signature.asc Description: Digital signature
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Julien Danjou wrote: At 1241861577 time_t, Uli Schlachter wrote: I have no idea what caused this and it doesn't seem reproducible which is why I won't open a bug report. I'm updating to rc2 now and I will run awesome in valgrind. Let's see if I can find anything... What I can suggest is to call ulimit -c unlimited before running awesome. At least you'll have a coredump. And if you can reproduce it, recompiling with -O0 -fno-inline can help to have a more complete bt. Cheers, Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( ==22611== Invalid read of size 1 ==22611==at 0x4A09F9E: memmove (mc_replace_strmem.c:517) ==22611==by 0x439E24: (within /usr/bin/awesome) ==22611==by 0x430031: (within /usr/bin/awesome) ==22611==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x41A843: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611== Address 0x8148ed8 is 0 bytes after a block of size 96 alloc'd ==22611==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==22611==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==22611==by 0x41A923: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611==by 0x40DFDC: (within /usr/bin/awesome) ==22611==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==22611==by 0x40E8A6: main (in /usr/bin/awesome) ==22611== ==22611== Invalid read of size 1 ==22611==at 0x4A09F90: memmove (mc_replace_strmem.c:517) ==22611==by 0x439E24: (within /usr/bin/awesome) ==22611==by 0x430031: (within /usr/bin/awesome) ==22611==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x41A843: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611== Address 0x8148ed9 is 1 bytes after a block of size 96 alloc'd ==22611==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==22611==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==22611==by 0x41A923: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611==by 0x40DFDC: (within /usr/bin/awesome) ==22611==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==22611==by 0x40E8A6: main (in /usr/bin/awesome) ==22611== ==22611== Invalid write of size 1 ==22611==at 0x4A09F94: memmove (mc_replace_strmem.c:517) ==22611==by 0x439E24: (within /usr/bin/awesome) ==22611==by 0x430031: (within /usr/bin/awesome) ==22611==by 0x30AAE0CB15: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D028: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE173DB: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0D084: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C706: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE0C784: (within /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x30AAE08093: lua_pcall (in /usr/lib/liblua5.1.so.0.0.0) ==22611==by 0x41A843: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611== Address 0x8148ed8 is 0 bytes after a block of size 96 alloc'd ==22611==at 0x4A0891E: malloc (vg_replace_malloc.c:207) ==22611==by 0x4A08AA7: realloc (vg_replace_malloc.c:429) ==22611==by 0x41A923: (within /usr/bin/awesome) ==22611==by 0x3565A00BE8: (within /usr/lib/libxcb-property.so.1.0.0) ==22611==by 0x40DFDC: (within /usr/bin/awesome) ==22611==by 0x32AE607216: ev_loop (in /usr/lib/libev.so.3.0.0) ==22611==by 0x40E8A6: main (in /usr/bin/awesome) ==22611== ==22611== More than 1000 total errors detected. I'm not reporting any more. ==22611== Final error counts will be inaccurate. Go fix your program! ==22611== Rerun with --error-limit=no to disable this cutoff. Note ==22611== that errors may occur in your program without prior warning from ==22611==
Re: 3.3-rc1 crash bug (?)
At 1241874772 time_t, Uli Schlachter wrote: Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( http://naquadah.org/~jd/debian/awesome_3.3~rc2-1_nostrip-debug-noopt_amd64.deb -- Julien Danjou // ᐰ jul...@danjou.info http://julien.danjou.info // 9A0D 5FD9 EB42 22F6 8974 C95C A462 B51E C2FE E5CD // And thinking so much differently. signature.asc Description: Digital signature
Re: 3.3-rc1 crash bug (?)
Uli Schlachter pisze: Julien Danjou wrote: At 1241861577 time_t, Uli Schlachter wrote: I have no idea what caused this and it doesn't seem reproducible which is why I won't open a bug report. I'm updating to rc2 now and I will run awesome in valgrind. Let's see if I can find anything... What I can suggest is to call ulimit -c unlimited before running awesome. At least you'll have a coredump. And if you can reproduce it, recompiling with -O0 -fno-inline can help to have a more complete bt. Cheers, Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( Hi, Do you have git version of xcb-proto ? If yes revert commit fe7b12db4fc0e95f3eef038581bf2154e1727c7a, Correct the length calculation for the value field of GetPropertyReply. ... and rebuild libxcb and xcb-util ... :p Mariusz Ceier -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Mariusz Ceier wrote: Uli Schlachter pisze: Julien Danjou wrote: At 1241861577 time_t, Uli Schlachter wrote: I have no idea what caused this and it doesn't seem reproducible which is why I won't open a bug report. I'm updating to rc2 now and I will run awesome in valgrind. Let's see if I can find anything... What I can suggest is to call ulimit -c unlimited before running awesome. At least you'll have a coredump. And if you can reproduce it, recompiling with -O0 -fno-inline can help to have a more complete bt. Cheers, Well, I went with valgrind and now got this backtrace (man, I miss debugging symbols... btw I still can't compile awesome due to some missing dependencies). Anyone got some ideas what's going on? Proposals how to get my beloved debug symbols? :( Hi, Do you have git version of xcb-proto ? If yes revert commit fe7b12db4fc0e95f3eef038581bf2154e1727c7a, Correct the length calculation for the value field of GetPropertyReply. ... and rebuild libxcb and xcb-util ... :p Mariusz Ceier Nope, this debian testing + awesome 3.3* from unstable (which pulled in libxcb-atom1, libxcb-keysyms1 and libxdg-basedir1 from unstable). jd: I guess the debian version doesn't have this... bug? - -- Do you know that books smell like nutmeg or some spice from a foreign land? -- Faber in Fahrenheit 451 -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEcBAEBCAAGBQJKBYjPAAoJECLkKOvLj8sG7PsIAIcCwC5mHNZcpTT+K4amAWVI AfWqK9uWonDcB6ZRJlm19cCoTNb7x56aXXZzyUJwbDxDPhEk7e8UbtTEIGVuZ2JU ObfZczyEmgdIe522nxLpYemSlX8USZLPqSGaxpfCF032CVfSFB7Ezgh7gav/gfZb ijuq+xNf1qChjFJom1QuThMTuLnd8px+fcUEEAqOMu7UpH97AAAlKGJjaw4BAh+p 889QO+lSMyf1Tm1cRSWwC04yOeI7Iil525AagLMLAMpwqnrEQ+8DQ7NPVVXqMo7H hrJvgQoQS+XXNT5d3iXkJoKIA2qdj5tcq8ZOln21eb9g5j777ci8xWztnxRDsmE= =G88D -END PGP SIGNATURE- -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
At 1241876690 time_t, Uli Schlachter wrote: jd: I guess the debian version doesn't have this... bug? Should not. -- Julien Danjou // ᐰ jul...@danjou.info http://julien.danjou.info // 9A0D 5FD9 EB42 22F6 8974 C95C A462 B51E C2FE E5CD // My root password is signature.asc Description: Digital signature
Re: 3.3-rc1 crash bug (?)
Do you have git version of xcb-proto ? If yes revert commit fe7b12db4fc0e95f3eef038581bf2154e1727c7a, Correct the length calculation for the value field of GetPropertyReply. ... and rebuild libxcb and xcb-util ... :p Shouldn't you get it reverted if it's a bad commit? -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.
Re: 3.3-rc1 crash bug (?)
Maarten Maathuis pisze: Do you have git version of xcb-proto ? If yes revert commit fe7b12db4fc0e95f3eef038581bf2154e1727c7a, Correct the length calculation for the value field of GetPropertyReply. ... and rebuild libxcb and xcb-util ... :p Shouldn't you get it reverted if it's a bad commit? I think I should ... bugreport: https://bugs.freedesktop.org/show_bug.cgi?id=21644 Mariusz Ceier -- To unsubscribe, send mail to awesome-devel-unsubscr...@naquadah.org.