Version 0.6 crashes

Question

Version 0.6 crashes

accelerator74 opened this issue 8 years ago · comments

Accelerator commented 8 years ago

https://crash.limetech.org/w4mqs5axdskl

Attano · Answer 1 · Thu Mar 24 2016 02:40:30 GMT+0800 (China Standard Time)

Why did you close this? Have you fixed this or what?

Accelerator · Answer 2 · Thu Mar 24 2016 02:57:26 GMT+0800 (China Standard Time)

No. But version 0.6 still falls, but there is no crash dump.
This dump version 0.5.7...

Accelerator · Answer 3 · Thu Mar 24 2016 02:58:14 GMT+0800 (China Standard Time)

I will try to provide a crash dump version 0.6 tomorrow.

Accelerator · Answer 4 · Thu Mar 24 2016 16:17:19 GMT+0800 (China Standard Time)

Crash dump version 0.6:
https://crash.limetech.org/f6wpj7bm7xw3
Version 0.5.7 stable for me. 0.6 - crashes.

Accelerator · Answer 5 · Thu Mar 24 2016 20:26:03 GMT+0800 (China Standard Time)

And again: https://crash.limetech.org/rtp3wjtum5b5

Attano · Answer 6 · Thu Mar 24 2016 21:13:56 GMT+0800 (China Standard Time)

What Linux distribution and version are you using?

Accelerator · Answer 7 · Thu Mar 24 2016 21:34:35 GMT+0800 (China Standard Time)

Your source code compiled on Ubuntu 14.04, gcc-4.8.

Accelerator · Answer 8 · Thu Mar 24 2016 21:34:53 GMT+0800 (China Standard Time)

Latest branches sourcemod and metamod

Accelerator · Answer 9 · Thu Mar 24 2016 21:36:59 GMT+0800 (China Standard Time)

Server OS: Linux Gentoo

Attano · Answer 10 · Thu Mar 24 2016 21:57:49 GMT+0800 (China Standard Time)

You can switch to Ubuntu 12.04 or newer. This is guaranteed to solve your issue.

Accelerator · Answer 11 · Thu Mar 24 2016 22:34:22 GMT+0800 (China Standard Time)

I found fix!
l4d2_unlock_melees 1
I used extension l4d2_meleespawncontrol. And the set cvar to 1 stopped crashes

Attano · Answer 12 · Fri Mar 25 2016 01:15:38 GMT+0800 (China Standard Time)

Nice to hear that!

Accelerator · Answer 13 · Thu Mar 31 2016 02:52:29 GMT+0800 (China Standard Time)

Crashes again..
I removed your changes - New forwards. And compiled without detours/inferno_spread.cpp and detours/shoved_by_pounce_landing.cpp. So, server full stable.

Matthew Ramir · Answer 14 · Tue Apr 12 2016 08:26:44 GMT+0800 (China Standard Time)

I also have this issue. @accelerator74 are you able to make a PR?

Matthew Ramir · Answer 15 · Tue Apr 12 2016 09:01:59 GMT+0800 (China Standard Time)

According to my disassembly the issue is with
(unsigned char *)target + 588) != 3
in inferno_spread.cpp

Unfortunately I'm not familiar enough to know what that offset should be.

Attano · Answer 16 · Tue Apr 12 2016 20:36:43 GMT+0800 (China Standard Time)

@crasx, that detour is not used anywhere, it is safe to remove it completely. Or you can wait for the next version, I'll do it myself.