We had a battle 3 vs 3 and all (except 2 players) crashed at the same time.
I watched demo and it crashes too with the same error.
gdb dump (from demo watching):
http://pastebin.com/3mNyigmL
infolog.txt (from demo watching) and demo are attached.
You can reproduce this bug by watching demo till the end (32+ minute).
mass crash - spring 87.0
Moderator: Moderators
mass crash - spring 87.0
- Attachments
-
- 20120316_051113_DeltaSiegeDry Revolution v2_87.sdf
- demo
- (1.81 MiB) Downloaded 112 times
-
- infolog.txt
- infolog
- (40.6 KiB) Downloaded 114 times
- Forboding Angel
- Evolution RTS Developer
- Posts: 14673
- Joined: 17 Nov 2005, 02:43
Re: mass crash - spring 87.0
Edit: Wrong Answer
Last edited by Forboding Angel on 17 Mar 2012, 00:40, edited 1 time in total.
Re: mass crash - spring 87.0
to me it looks like a bad_alloc was caugh in an exception catch block, that should actually not have caught it. due to this, the required stack-trace information was lost, and is not visible in the GDB output.
it might be related to this commit:
https://github.com/spring/spring/commit ... 05b97b2456
it might be related to this commit:
https://github.com/spring/spring/commit ... 05b97b2456
Re: mass crash - spring 87.0
hoijui, could you please reproduce this crash yourself? using attached demo.
I am complete noob in gdb so I can miss some critical info. Maybe you can even debug this with debugger? I mean via breakpoint in ide.
I think that this is very important bug. Mass crash after 30 minute game - all players were disappointed and left battleroom.
I am complete noob in gdb so I can miss some critical info. Maybe you can even debug this with debugger? I mean via breakpoint in ide.
I think that this is very important bug. Mass crash after 30 minute game - all players were disappointed and left battleroom.
Re: mass crash - spring 87.0
The error was thrown in a non-mainthread thread. Also that gdb wasn't able to create is a known issue (we should add a gdb detection code (or copy it from watchdog) and rethrow the exception in such cases).
Also I am unsure if Spring doesn't print a stacktrace in this case itself (the infolog.txt pasted here is cut, either type a `sync` in the bash next time or copy the bash output)
Also I am unsure if Spring doesn't print a stacktrace in this case itself (the infolog.txt pasted here is cut, either type a `sync` in the bash next time or copy the bash output)
Re: mass crash - spring 87.0
jK, I attached with gdb to spring after crash. Next time I will do "sync" but I am sure that this will not help.
Please reproduce this bug using atached demo. You will be able to get all info you need.
Please reproduce this bug using atached demo. You will be able to get all info you need.
-
- Posts: 834
- Joined: 19 May 2009, 21:10
Re: mass crash - spring 87.0
SirMaverick, thanks!!!