View Issue Details

Jump to NotesJump to History
IDProjectCategoryView StatusDate SubmittedLast Update
0004279Spring engineLinuxpublic2014-01-18 23:142014-08-24 22:15
Reporterjamerlan Assigned Toabma  
PrioritynormalSeveritycrashReproducibilitysometimes
Status resolvedResolutionfixed 
Product Version96.0 
Summary0004279: crash after game end in~CVertexArray
Descriptioninfolog.txt (I compiled spring release so i have translated stack)

http://pastebin.com/yB580cZb
TagsNo tags attached.
Checked infolog.txt for Errors

Relationships

related to 0004282 resolvedjK another endgame crash with bt full 
related to 0004281 closedjK spring 96 is frozen on exit of the game 

Activities

cleanrock

2014-01-19 05:48

reporter   ~0012722

Last edited: 2014-01-19 06:39

 Can you attach the demo or a gdb backtrace (bt full) ?

jamerlan

2014-01-19 10:07

reporter   ~0012723

this is debug build so stack trace is translated. Or I am wrong?
Can't find demo file :(


[f=0079672] Error: Aborted (SIGABRT) in spring 96.0
[f=0079672] Error: Stacktrace for Spring 96.0:
[f=0079672] Error: <0> sigaction.c:?
[f=0079672] Error: <1> :?
[f=0079672] Error: <2> :?
[f=0079672] Error: <3> :?
[f=0079672] Error: <4> malloc.c:?
[f=0079672] Error: <5> malloc.c:?
[f=0079672] Error: <6> /usr/include/c++/4.8.2/bits/stl_iterator.h:748
[f=0079672] Error: <7> /home/revenant/projects/spring96/spring/rts/Game/UI/EndGameBox.h:55
[f=0079672] Error: <8> /usr/include/c++/4.8.2/bits/stl_vector.h:646
[f=0079672] Error: <9> /home/revenant/projects/spring96/spring/rts/System/float3.h:207
[f=0079672] Error: <10> /home/revenant/projects/spring96/spring/rts/System/Util.h:176
[f=0079672] Error: <11> /home/revenant/projects/spring96/spring/rts/Game/Game.cpp:523
[f=0079672] Error: <12> /usr/include/c++/4.8.2/bits/stl_map.h:484
[f=0079672] Error: <13> /usr/include/boost/smart_ptr/detail/shared_count.hpp:417
[f=0079672] Error: <14> /usr/include/c++/4.8.2/bits/basic_string.h:716
[f=0079672] Error: <15> crtstuff.c:?
[f=0079672] Error: <16> ??:?
[f=0079672] Error: <17> ??:114 (discriminator 1)
[f=0079672] Error: [ErrorMessageBox][1] msg="Spring has crashed:

jamerlan

2014-01-19 10:08

reporter   ~0012724

when battle ended - lot's of people got crash. For example: kaon (BD) got crash too

cleanrock

2014-01-19 11:02

reporter   ~0012725

infolog stacktrace do not make much sense, i think they are always pretty useless atm.
Since i see this in infolog
[f=0079672] [~CGame][8]
[f=0079672] Error: Aborted (SIGABRT) in spring 96.0
my conclusion is that crash happen on one of these 3 lines in ~CGame:
LOG("[%s][8]", __FUNCTION__);
SafeDelete(featureHandler); // depends on unitHandler (via ~CFeature)
SafeDelete(unitHandler); // depends on modelParser (via ~CUnit)
SafeDelete(projectileHandler);
LOG("[%s][9]", __FUNCTION__);

silentwings

2014-01-19 11:26

reporter   ~0012726

Last edited: 2014-01-19 11:27

 Does the demo crash when you watch it? If so please link to it (and I'll get a gdb trace from it).

jamerlan

2014-01-19 11:59

reporter   ~0012729

silentwings looks like demo was not written

cleanrock maybe code review can help. I mean looks like bug is some where here:
[f=0079672] Error: <7> /home/revenant/projects/spring96/spring/rts/Game/UI/EndGameBox.h:55

silentwings

2014-01-19 12:02

reporter   ~0012730

Did you check the replay site? replays.springrts.com

jamerlan

2014-01-19 12:47

reporter   ~0012732

Last edited: 2014-01-19 12:47

 >>Did you check the replay site? replays.springrts.com
I was spectator(I think, don't remember) and I have no idea how to find that battle

cleanrock

2014-01-19 12:47

reporter   ~0012733

Last edited: 2014-01-19 12:48

 jamerlan,
I looked at EndGameBox.h:55 first, i doubt problem is there, infolog stacktraces are crap.
We should try to get our hands on a replay that crashes the same way, ur infolog says this:
[f=0000000] recording demo: demos/20140119_012420_DeltaSiegeDry_96.sdf
Can you find that file on you computer ?

jamerlan

2014-01-19 12:50

reporter   ~0012734

>Can you find that file on you computer ?
no it does not exists

silentwings

2014-01-19 12:53

reporter   ~0012735

So look on replays.springrts.com for a replay matching that map/time/date/game/host?

jamerlan

2014-01-19 14:09

reporter   ~0012738

about that:

>infolog stacktrace do not make much sense, i think they are always pretty >useless atm.
>Since i see this in infolog
>[f=0079672] [~CGame][8]
>[f=0079672] Error: Aborted (SIGABRT) in spring 96.0
>my conclusion is that crash happen on one of these 3 lines in ~CGame:
>LOG("[%s][8]", __FUNCTION__);
>SafeDelete(featureHandler); // depends on unitHandler (via ~CFeature)
>SafeDelete(unitHandler); // depends on modelParser (via ~CUnit)
>SafeDelete(projectileHandler);
>LOG("[%s][9]", __FUNCTION__);


cleanrock, in Game.cpp:523
I have this:
if (!gu->globalQuit) PostLoadRendering();

and in EndGameBox.h:55
I have: if (team >= 0 && static_cast<size_t>(team) >= values.size()) {

cleanrock

2014-01-19 14:18

reporter   ~0012739

jamerlan,
Game.cpp:523 -> if (!gu->globalQuit) PostLoadRendering();
do not make sense at all.
Are you looking at 96.0 source ?
My Game.cpp:523 is end (end brace) in CGame dtor.
I dont trust infolog stacktraces a bit.
It would be very good if we could have proper stacktraces in infolog but i think its very difficult to achieve that.

abma

2014-01-20 18:32

administrator   ~0012746

i wonder a bit about the signal,

why is it SIGABRT?

cleanrock

2014-01-20 20:17

reporter   ~0012747

Last edited: 2014-01-20 20:19

 Good point, assert prints dont end up in infolog right ? If they dont perhaps we can make them be printed there as well.
jamerlan says its a debug build, only asserts i see is within
CProjectile::Detach()
called from the SafeDelete(projectileHandler); in ~CGame

abma

2014-01-20 22:44

administrator   ~0012749

the infolog.txt says its not a debug build, so asserts would be disabled...

abma

2014-08-01 16:50

administrator   ~0013459

does this still happen with current development version / spring 97?

jamerlan

2014-08-05 19:32

reporter   ~0013472

will try to reproduce with replay for spring 96. For me it crashes(freezes at the end actually) from time to time. And then I kill game with kill -9.

Looks like it only happens in massive and long multiplayer games. So I can't currently reproduce it in spring 97 (latest development)

I will try more at the weekends (takes lots of time)

jamerlan

2014-08-09 21:23

reporter   ~0013498

Spring 96 freezed at the end of the game and I did "bt full" using gdb:
http://pastebin.com/rSzDvSYs

cleanrock

2014-08-09 22:19

reporter   ~0013499

This non-critical bug is not of interest in 96.0 imo. Please check if the bug is in latest develop (97.0.1).

jamerlan

2014-08-09 22:28

reporter   ~0013500

This bug is only reproducible if you are playing mass team game as a player (not a spec) in game end. I can't test it in spring 97 because no mass games

abma

2014-08-22 21:57

administrator   ~0013553

sorry, i don't see how we could fix this bug, as its at the end + rarely it seems its not a very big problem.

98.0 will contain A LOT of changes, so lets see if it crashes then, too.

also atm i don't trust the linux crashhandler, the stacktraces are mostly broken :-|

abma

2014-08-22 22:18

administrator   ~0013554

oops, didn't see your backtrace, sorry.

for me it looks like sth. is wrong in rts/Rendering/Env/Decals/GroundDecalHandler.cpp:124, it crashes because a scar is deleted twice.

cleanrock

2014-08-24 09:20

reporter   ~0013559

Crash in ~CVertexArray so this may already be fixed by https://github.com/spring/spring/commit/708c3d6

abma

2014-08-24 22:15

administrator   ~0013564

indeed, thanks!

Issue History

Date Modified Username Field Change
2014-01-18 23:14 jamerlan New Issue
2014-01-19 05:48 cleanrock Note Added: 0012722
2014-01-19 06:39 cleanrock Note Edited: 0012722
2014-01-19 10:07 jamerlan Note Added: 0012723
2014-01-19 10:08 jamerlan Note Added: 0012724
2014-01-19 11:02 cleanrock Note Added: 0012725
2014-01-19 11:26 silentwings Note Added: 0012726
2014-01-19 11:27 silentwings Note Edited: 0012726
2014-01-19 11:27 silentwings Note Edited: 0012726
2014-01-19 11:59 jamerlan Note Added: 0012729
2014-01-19 12:02 silentwings Note Added: 0012730
2014-01-19 12:47 jamerlan Note Added: 0012732
2014-01-19 12:47 cleanrock Note Added: 0012733
2014-01-19 12:47 jamerlan Note Edited: 0012732
2014-01-19 12:48 cleanrock Note Edited: 0012733
2014-01-19 12:50 jamerlan Note Added: 0012734
2014-01-19 12:53 silentwings Note Added: 0012735
2014-01-19 14:09 jamerlan Note Added: 0012738
2014-01-19 14:18 cleanrock Note Added: 0012739
2014-01-20 15:54 abma Relationship added related to 0004281
2014-01-20 15:54 abma Relationship added related to 0004282
2014-01-20 18:32 abma Note Added: 0012746
2014-01-20 20:17 cleanrock Note Added: 0012747
2014-01-20 20:19 cleanrock Note Edited: 0012747
2014-01-20 22:44 abma Note Added: 0012749
2014-01-21 12:42 abma Severity minor => crash
2014-08-01 16:50 abma Note Added: 0013459
2014-08-01 16:50 abma Status new => feedback
2014-08-05 19:32 jamerlan Note Added: 0013472
2014-08-05 19:32 jamerlan Status feedback => new
2014-08-09 21:23 jamerlan Note Added: 0013498
2014-08-09 22:19 cleanrock Note Added: 0013499
2014-08-09 22:23 cleanrock Status new => feedback
2014-08-09 22:28 jamerlan Note Added: 0013500
2014-08-09 22:28 jamerlan Status feedback => new
2014-08-22 21:57 abma Note Added: 0013553
2014-08-22 21:57 abma Status new => closed
2014-08-22 21:57 abma Assigned To => abma
2014-08-22 21:57 abma Resolution open => unable to reproduce
2014-08-22 22:18 abma Note Added: 0013554
2014-08-22 22:18 abma Assigned To abma =>
2014-08-22 22:18 abma Status closed => new
2014-08-24 09:20 cleanrock Note Added: 0013559
2014-08-24 22:14 abma Summary crash after game end => crash after game end in~CVertexArray
2014-08-24 22:15 abma Note Added: 0013564
2014-08-24 22:15 abma Resolution unable to reproduce => fixed
2014-08-24 22:15 abma Status new => resolved
2014-08-24 22:15 abma Assigned To => abma