View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0005665 | Spring engine | General | public | 2017-07-21 21:52 | 2017-08-19 14:34 |
| Reporter | Jools | Assigned To | |||
| Priority | normal | Severity | crash | Reproducibility | always |
| Status | closed | Resolution | reopened | ||
| Product Version | 103.0 +git | ||||
| Summary | 0005665: Crash while loading: rts/Rendering/Models/3DModel.cpp:76 | ||||
| Description | https://github.com/spring/spring/tree/103.0.1-1225-g87b33dd ??:0 ??:0 ??:0 ??:0 rts/Rendering/Models/3DModel.cpp:76 rts/Rendering/Models/IModelParser.cpp:313 rts/Rendering/Models/IModelParser.cpp:261 rts/Rendering/Models/IModelParser.cpp:227 rts/Sim/Objects/SolidObjectDef.cpp:81 rts/Lua/LuaSyncedRead.cpp:3739 rts/lib/lua/src/ldo.cpp:321 rts/lib/lua/src/lvm.cpp:112 rts/lib/lua/src/lvm.cpp:613 /opt/mxe/tmp-gcc/gcc-4.8.1.build/i686-w64-mingw32/libgcc/./gthr-default.h:628 rts/lib/lua/src/ldo.cpp:378 /opt/mxe/tmp-gcc/gcc-4.8.1.build/i686-w64-mingw32/libgcc/./gthr-default.h:628 rts/lib/lua/src/lapi.cpp:813 rts/lib/lua/src/ldo.cpp:119 /opt/mxe/tmp-gcc/gcc-4.8.1/libstdc++-v3/src/c++11/chrono.cc:76 rts/lib/lua/src/ldo.cpp:465 rts/System/Misc/SpringTime.h:35 rts/lib/lua/src/lbaselib.cpp:378 rts/lib/lua/src/ldo.cpp:321 rts/lib/lua/src/lvm.cpp:613 /opt/mxe/tmp-gcc/gcc-4.8.1.build/i686-w64-mingw32/libgcc/./gthr-default.h:628 ??:0 ??:0 rts/lib/mingw-std-threads/mingw.thread.h:157 /opt/mxe/usr/lib/gcc/i686-w64-mingw32/4.8.1/include/c++/functional:2073 /opt/mxe/usr/lib/gcc/i686-w64-mingw32/4.8.1/include/c++/functional:2029 ??:0 ??:0 ??:0 ??:0 | ||||
| Additional Information | http://paste.springfiles.com/view/badb2a71 | ||||
| Tags | No tags attached. | ||||
| Checked infolog.txt for Errors | |||||
|
|
It's related to using two monitors somehow: it doesn't crash when I just use one screen |
|
|
it's an ATI driver bug, out of our control. see also 0005643 0005586 0005527 0005484. |
|
|
I have Nvidia, how can it be related to ATI driver? |
|
|
All other instances of this crash were from ATI users, you're the first to break that pattern. This is why infologs exist. I don't have two displays to verify what's going on here but I doubt it can be fixed engine-side. |
|
|
http://paste.springfiles.com/view/f44def2f Here's one. Seems to be quite reproducible with the two monitor setting, but I think it worked once even with two monitors enabled. |
|
|
Does it reproduce if you add Fullscreen=0 to springsettings? Also, with fullscreen enabled, add DebugGL=1 and DebugGLStacktraces=1 to find out if the driver has anything to say before it crashes. |
|
|
Can't reproduce it anymore: it started working also with two screens. (With DebugGL=1 the game freezes on start in a black screen, waited for some minutes before i killed the process.) It could be something else or it could be related with when the user enters two screen (extend desktop) mode, in windows I frequently change modes with the win+P keystroke. |
|
|
hmm, non-determinism in system behavior doesn't sound promising. if you can definitively link these crashes to an action in Windows (and they haven't been eliminated in later driver versions; 347.5 is a few years old) you might want to report this to nvidia too. |
|
|
Just reopening to add some comments, feel free to close... It's not that the system isn't deterministic, it's just that there are so many ways that the user can toggle screen modes. For instance: - starting the game in dual screen mode but using single screen mode conf file - going to dual screen mode while game is loading - using dual screen mode but focus on the other screen - etc Also, it could be that once something is correctly cached it will no longer crash even if the same conditions persist. I don't believe in non-deterministic computers |
|
|
I deliberately downgraded the nvidia driver because the most recent one became unstable on my windows 7 system. |
|
|
Modern systems (hardware and software) are so complicated that from the perspective of a program their operation might as well be non-deterministic; the internal state has grown far too large and opaque to even try reasoning about. Do enough development and you slowly learn to accept this. I still have no clue what the theoretical connection between this crash (in either a glNewList or glGenLists call) and extended desktop mode should even be. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2017-07-21 21:52 | Jools | New Issue | |
| 2017-07-21 21:56 | Jools | Note Added: 0018089 | |
| 2017-07-21 22:11 | Kloot | Note Added: 0018090 | |
| 2017-07-21 22:13 | Jools | Note Added: 0018091 | |
| 2017-07-21 22:50 | Kloot | Note Added: 0018092 | |
| 2017-07-21 22:54 | Kloot | Note Edited: 0018092 | |
| 2017-07-21 22:58 | Jools | Note Added: 0018093 | |
| 2017-07-21 23:16 | Kloot | Note Added: 0018094 | |
| 2017-07-21 23:17 | Kloot | Note Edited: 0018094 | |
| 2017-07-21 23:49 | Jools | Note Added: 0018095 | |
| 2017-07-22 00:46 | Kloot | Status | new => closed |
| 2017-07-22 00:46 | Kloot | Resolution | open => unable to reproduce |
| 2017-07-22 00:46 | Kloot | Note Added: 0018096 | |
| 2017-07-22 12:17 | Jools | Status | closed => feedback |
| 2017-07-22 12:17 | Jools | Resolution | unable to reproduce => reopened |
| 2017-07-22 12:17 | Jools | Note Added: 0018106 | |
| 2017-07-22 12:18 | Jools | Note Added: 0018107 | |
| 2017-07-22 12:18 | Jools | Status | feedback => new |
| 2017-07-22 13:34 | Kloot | Note Added: 0018108 | |
| 2017-07-22 13:35 | Kloot | Note Edited: 0018108 | |
| 2017-08-19 14:34 | Kloot | Status | new => closed |
