I downloaded and compiled the source provided on hoijui's "Engine Testing" thread hoping that a crash I was experiencing in the stable 0.82.7.1 multithreaded version of spring had been fixed, but it seems I am still experiencing the same issue. After playing BA v7.50 for a period of time (I have crashed in as little as 30 minutes & made it as long as 2+ hours), the game will just freeze with no slowdown or prior warning. I have only played against the AI & the freeze usually seems to occur when I launch a large attack & start shipping my units across the map. It only occurs in the MT version of spring as I have made it through many games using the normal version (albeit at a MUCH lower frame rate). Perhaps I am just pushing the engine too far & allowing too many units onto the board, but it seems to be somewhat reproducible so I thought I'd try posting here. I have attached my infolog.txt as requested in the error dialog & I'll try to help in any way that I can. (And yes, if you look at the infolog, i used /atm a few times...still a NOOB)
OS: Ubuntu 11.04 (natty), 64 bit, 2.6.38-10-generic, Fully Updated
Graphics: GeForce GTS 250 using NVIDIA's 270.41.06 Driver
CPU: AMD Phenom II X4 Processor
RAM: 2 GB
SIGSEGV in spring 0.82+.4.0 (MT-Sim)
Moderator: Moderators
SIGSEGV in spring 0.82+.4.0 (MT-Sim)
- Attachments
-
- infolog.txt
- infolog.txt
- (32.13 KiB) Downloaded 98 times
Re: SIGSEGV in spring 0.82+.4.0 (MT-Sim)
thanks! three wishes to make it more useful:
1.
compile with debug3 (this will really slow it down!) mode and signal_nans=true.
the stacktrace output from this infolog.txt doesn't help, sorry. it looks like stack/memory corruption.
2. add the output of
"git describe" maybe this is already fixed. currently most commits are fixes! :)
3.
report to mantis: http://springrts.com/mantis/
1.
compile with debug3 (this will really slow it down!) mode and signal_nans=true.
Code: Select all
cmake -DSIGNAL_NANS=TRUE -DCMAKE_BUILD_TYPE=DEBUG3
2. add the output of
"git describe" maybe this is already fixed. currently most commits are fixes! :)
3.
report to mantis: http://springrts.com/mantis/
Re: SIGSEGV in spring 0.82+.4.0 (MT-Sim)
Thanks much for the reply. Maybe I'll recompile the latest master with your instructions here in the next day or two & see if I still have the problem. Spring is great & keep up the good work :)
Re: SIGSEGV in spring 0.82+.4.0 (MT-Sim)
as a note: when compiled with signal-nans, you've to run it with gdb:
[...]
if it crashes type:
without that, it could simply quit with a SIGFPE message and it won't do any usefull output.
at best, you can give instructions how to easily reproduce the crash/desync/...
some spring commands that could help:
/cheat
/give all
/give all 1
/give 10 armcom
/nocost
and maybe https://github.com/spring/testmonkey is useful, too...
Code: Select all
gdb /path/to/spring
run
if it crashes type:
Code: Select all
bt full
at best, you can give instructions how to easily reproduce the crash/desync/...
some spring commands that could help:
/cheat
/give all
/give all 1
/give 10 armcom
/nocost
and maybe https://github.com/spring/testmonkey is useful, too...
Re: SIGSEGV in spring 0.82+.4.0 (MT-Sim)
ok, thanks again. I don't have much experience debugging things, so any info I can get is helpful. 
