[asherkin]

First is (just look at the stack trace), 2nd is a memory corruption issue.

[4bdul]

I believe I am getting identical crashes - also Windows OS. It also happens when map changes.

https://crash.limetech.org/p4q2trs2zysb
https://crash.limetech.org/ykbwybayzvyh
https://crash.limetech.org/va66by3koevl

I don't have the movement unlocker plugin.

[asherkin]

The crash is deep in the map loading code, when it's loading models - the function at the top of the stack is CModelLoader::Studio_LoadModel.
My guess from where it is crashing is that you've got a corrupt model.

[dilalmon]

I've recently found out that this has nothing to do with Sourcemod. Sorry for misleading everyone.
When I disabled Movement Unlocker, it ran for a very long time without crashing. My latest test shows this is false.
I disabled Sourcemod entirely and it was still crashing.

I was also able to recreate it. I just had to load some maps a few times (changelevel).

If it was due to a corrupt model, assuming from a map, why would it load without a problem previously?

I removed -nocrashdialog and this is only seen right before the crash, and never when it loads fine:



Any ideas, asherkin?

[Nolongerinthegame]

Originally Posted by dilalmon I've recently found out that this has nothing to do with Sourcemod. Sorry for misleading everyone. When I disabled Movement Unlocker, it ran for a very long time without crashing. My latest test shows this is false. I disabled Sourcemod entirely and it was still crashing. I was also able to recreate it. I just had to load some maps a few times (changelevel). If it was due to a corrupt model, assuming from a map, why would it load without a problem previously? I removed -nocrashdialog and this is only seen right before the crash, and never when it loads fine: Any ideas, asherkin?

Interesting that it crashes without Sourcemod too.

[Potato Uno]

Originally Posted by asherkin First is (just look at the stack trace), 2nd is a memory corruption issue.

How are you able to make sense out of Windows stack traces when they have no symbols?

TELL US YOUR SECRETS!

[Dr. Greg House]

Time and sweat.

[asherkin]

Originally Posted by Potato Uno How are you able to make sense out of Windows stack traces when they have no symbols? TELL US YOUR SECRETS!

They have symbols, just no non-SourceMod-or-Windows-OS symbols.

The first was a few frames after the PrecacheModel native, which I know from the distance and the fact it's in engine.dll that the function in question is Sys_Error. For the second, the "crash" is triggered by the CRT when it detects heap corruption (look at the function names).

For other crashes in Valveland, it's just reverse signature finding - find the matching function at the given address in the Linux binaries.

[Nolongerinthegame]

I've noticed not just mine but others posts about these precache model crashes. They have been a lot more common recently, I wonder if its any update-related by Valve in December.