0003200: crashhandler / translated stacktrace seems to be broken

Notes
(0009011) abma (administrator) 2012-07-16 01:01 edited on: 2012-07-16 01:03	/cheat /crash leads to: C:\Games\Spring\spring.exe 0x004DFE08 rts/Game/UnsyncedGameCommands.cpp:2877 C:\Games\Spring\spring.exe 0x0042F7CF rts/Game/IActionExecutor.h:120 C:\Games\Spring\spring.exe 0x0042F948 rts/Game/Game.cpp:2327 C:\Games\Spring\spring.exe 0x004E8FA2 rts/Game/UnsyncedGameCommands.cpp:118 C:\Games\Spring\spring.exe 0x004EA749 rts/Game/UnsyncedGameCommands.cpp:143 C:\Games\Spring\spring.exe 0x0042FB83 rts/Game/Game.cpp:782 C:\Games\Spring\spring.exe 0x00935D74 rts/System/SpringApp.cpp:1183 C:\Games\Spring\spring.exe 0x00D6CA60 /slave/mingwlibs/include/boost/signals/detail/slot_call_iterator.hpp:61 C:\Games\Spring\spring.exe 0x00CADDB7 /slave/mingwlibs/include/boost/optional/optional.hpp:438 C:\Games\Spring\spring.exe 0x0093D03F rts/System/SpringApp.cpp:1011 C:\Games\Spring\spring.exe 0x00923761 rts/System/Main.cpp:51 C:\Games\Spring\spring.exe 0x00923CC4 rts/System/Main.cpp:65 C:\Games\Spring\spring.exe 0x00923DE1 rts/System/Main.cpp:101 C:\Games\Spring\spring.exe 0x009C15B9 GameData.cpp:0 C:\Games\Spring\spring.exe 0x004010B4 /root/mingw/mingw32-runtime-3.15.2/build_dir/src/mingwrt-3.15.2-mingw32/crt1.c:237 C:\Games\Spring\spring.exe 0x00401128 /root/mingw/mingw32-runtime-3.15.2/build_dir/src/mingwrt-3.15.2-mingw32/crt1.c:271 C:\WINDOWS\system32\kernel32.dll 0x7C817077 ??:0 that really looks broken... GameData.cpp:0 ???

(0009013) abma (administrator) 2012-07-16 01:27	on linux it looks better: f=0001708] Cheating is enabled! [f=0001804] Error: Segmentation fault (SIGSEGV) in spring 89.0 (OMP) [f=0001804] Error: Stacktrace: [f=0001804] Error: <0> [0x618400] [f=0001804] Error: <1> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/UnsyncedGameCommands.cpp:2877 [f=0001804] Error: <2> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/IActionExecutor.h:120 [f=0001804] Error: <3> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/Game.cpp:2327 [f=0001804] Error: <4> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/Action.h:8 [f=0001804] Error: <5> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/UnsyncedGameCommands.cpp:143 [f=0001804] Error: <6> /build/buildd/spring-89.0~12.04~ppa1/rts/Game/Game.cpp:782 [f=0001804] Error: <7> /build/buildd/spring-89.0~12.04~ppa1/rts/System/SpringApp.cpp:1183 [f=0001804] Error: <8> /usr/include/c++/4.6/bits/stl_list.h:154 [f=0001804] Error: <9> /build/buildd/spring-89.0~12.04~ppa1/rts/System/SpringApp.cpp:1011 [f=0001804] Path cache hits 0 0%

(0009015) zerver (developer) 2012-07-16 03:22	Is it really so broken? Those traces are fairly similar. AFAIK only debug builds are somewhat guaranteed to be able to unwind the stack. In release builds it only gives an impression of what the real stack trace looks like.

(0009016) abma (administrator) 2012-07-16 07:59	thats how it looks like with gdb: bt #0 (anonymous namespace)::CrashActionExecutor::Execute (this=0xc63ca78, action=...) at /home/matze/dev/spring/rts/Game/UnsyncedGameCommands.cpp:2877 #1 0x080e5700 in ExecuteAction (action=..., this=<optimized out>) at /home/matze/dev/spring/rts/Game/IActionExecutor.h:120 0000002 CGame::ActionPressed (this=0x927d358, key=13, action=..., isRepeat=false) at /home/matze/dev/spring/rts/Game/Game.cpp:2361 0000003 0x080e5855 in CGame::ProcessAction (this=0x927d358, action=..., key=13, isRepeat=false) at /home/matze/dev/spring/rts/Game/Game.cpp:2327 0000004 0x081cca16 in CGame::ProcessCommandText (this=0x927d358, key=13, command=...) at /home/matze/dev/spring/rts/Game/UnsyncedGameCommands.cpp:117 0000005 0x081ced2c in CGame::ProcessKeyPressAction (this=0x927d358, key=13, action=...) at /home/matze/dev/spring/rts/Game/UnsyncedGameCommands.cpp:143 0000006 0x080e5a90 in CGame::KeyPressed (this=0x927d358, key=17728, isRepeat=false) at /home/matze/dev/spring/rts/Game/Game.cpp:782 0000007 0x085f16d5 in SpringApp::MainEventHandler (this=0xbffff198, event=...) at /home/matze/dev/spring/rts/System/SpringApp.cpp:1182 0000008 0x085c57d0 in operator() (a0=..., this=<optimized out>) at /usr/include/boost/function/function_template.hpp:1013 0000009 operator()<boost::signals::detail::connection_slot_pair> (slot=..., this=<optimized out>) at /usr/include/boost/signals/signal_template.hpp:119 0000010 dereference (this=<optimized out>) at /usr/include/boost/signals/detail/slot_call_iterator.hpp:61 0000011 dereference<boost::signals::detail::slot_call_iterator<boost::signals::detail::call_bound1<void>::caller<const SDL_Event&, boost::function<void(const SDL_Event&)> >, boost::signals::detail::named_slot_map_iterator> > (f=...) at /usr/include/boost/iterator/iterator_facade.hpp:517 0000012 operator* (this=<optimized out>) at /usr/include/boost/iterator/iterator_facade.hpp:643 0000013 postfix_increment_proxy (x=..., this=<optimized out>) at /usr/include/boost/iterator/iterator_facade.hpp:145 0000014 operator++<boost::signals::detail::slot_call_iterator<boost::signals::detail::call_bound1<void>::caller<const SDL_Event&, boost::function<void(const SDL_Event&)> >, boost::signals::detail::named_slot_map_iterator>, boost::signals::detail::unusable, boost::single_pass_traversal_tag, const boost::signals::detail::unusable&, int> (i=...) at /usr/include/boost/iterator/iterator_facade.hpp:732 0000015 operator()<boost::signals::detail::slot_call_iterator<boost::signals::detail::call_bound1<void>::caller<const SDL_Event&, boost::function<void(const SDL_Event&)> >, boost::signals::detail::named_slot_map_iterator> > (last=..., first=..., this=<optimized out>) at /usr/include/boost/last_value.hpp:49 0000016 boost::signal1<void, SDL_Event const&, boost::last_value<void>, int, std::less<int>, boost::function<void (SDL_Event const&)> >::operator()(SDL_Event const&) (this=0x8cdb72c, a1=...) at /usr/include/boost/signals/signal_template.hpp:354 ---Type <return> to continue, or q <return> to quit--- 0000017 0x085f59e5 in SpringApp::Run (this=0xbffff198, argc=1, argv=0xbffff384) at /home/matze/dev/spring/rts/System/SpringApp.cpp:1014 0000018 0x085d3453 in MainFunc (argc=207866488, argv=0xc63ca78, ret=0x89f4540) at /home/matze/dev/spring/rts/System/Main.cpp:51 0000019 0x085d39c8 in Run (argc=1, argv=0xbffff384) at /home/matze/dev/spring/rts/System/Main.cpp:62 0000020 0x085d3abb in main (argc=1, argv=0xbffff384) at /home/matze/dev/spring/rts/System/Main.cpp:92

(0009017) abma (administrator) 2012-07-16 08:04 edited on: 2012-07-16 08:40	aaah, damit, it was limited to 10... why such a small limit? https://github.com/spring/spring/commit/29051ad82932f868c08ab7c317079c3488c52af3 [^]

(0009035) abma (administrator) 2012-07-16 19:27 edited on: 2012-07-16 19:38	ok, its broken in some cases imo: (thats locally, RELWITHDEBINFO, SIGNAN) http://paste.springfiles.com/view/raw/b6e8c67d [^] and thats a bt from the validation buildslave created by spring: (DEBUG2, SIGNAN) http://paste.springfiles.com/view/raw/12c3a5db [^] they should be the same! (but aren't....)

(0009036) abma (administrator) 2012-07-16 20:09	buildslave SIGNAN, RELWITHDEBINFO is broken, too: http://paste.springfiles.com/view/db5aa434 [^] so its not the buildtype, maybe its addr2line on the buildslave?!

(0009039) jK (developer) 2012-07-17 02:09	They shouldn't be the same. DEBUG2 disables nearly all optimizations and so the code is executed in a similar way it is written, while with RELEASE some massive optimizations are done which can restructure the code totally. Also RELWITHDEBINFO has holes in the debugsymbols (again cause of optimizations).

(0009043) abma (administrator) 2012-07-17 14:15	yep, but they should be similar. the bt's from the buildslave really looks broken: ( http://paste.springfiles.com/view/db5aa434 [^] ) so something is wrong at the buildslave, at my computer the bt's make sense...

(0009103) abma (administrator) 2012-08-03 20:59 edited on: 2012-08-03 21:16	hmm, investigated a bit (command & output of addr2line:) http://paste.springfiles.com/view/d38bd5e0 [^] vs gdb http://paste.springfiles.com/view/bd8bb7c7 [^] not sure whats broken on fedora with gcc 4.7... (thats on the validation buildlsave)

(0009104) abma (administrator) 2012-08-03 21:19	seems like some invalid addresses are included. every second (?) address seems to be invalid.

(0009115) abma (administrator) 2012-08-09 11:31	sorry, no idea how to fix :-/

Issue History
Date Modified	Username	Field	Change
2012-07-16 00:59	abma	New Issue
2012-07-16 01:01	abma	Note Added: 0009011
2012-07-16 01:02	abma	Note Edited: 0009011	View Revisions
2012-07-16 01:03	abma	Note Edited: 0009011	View Revisions
2012-07-16 01:03	abma	Relationship added	related to 0003193
2012-07-16 01:27	abma	Note Added: 0009013
2012-07-16 03:22	zerver	Note Added: 0009015
2012-07-16 07:59	abma	Note Added: 0009016
2012-07-16 08:04	abma	Note Added: 0009017
2012-07-16 08:40	abma	Note Edited: 0009017	View Revisions
2012-07-16 08:40	abma	Status	new => resolved
2012-07-16 08:40	abma	Resolution	open => fixed
2012-07-16 08:40	abma	Assigned To	=> abma
2012-07-16 19:27	abma	Note Added: 0009035
2012-07-16 19:27	abma	Status	resolved => new
2012-07-16 19:38	abma	Note Edited: 0009035	View Revisions
2012-07-16 20:09	abma	Note Added: 0009036
2012-07-17 02:09	jK	Note Added: 0009039
2012-07-17 14:15	abma	Note Added: 0009043
2012-08-03 20:59	abma	Note Added: 0009103
2012-08-03 20:59	abma	Note Edited: 0009103	View Revisions
2012-08-03 21:16	abma	Note Edited: 0009103	View Revisions
2012-08-03 21:19	abma	Note Added: 0009104
2012-08-09 11:31	abma	Assigned To	abma =>
2012-08-09 11:31	abma	Note Added: 0009115