Crash server DarkRP

[FunnyCraft]

https://pastebin.com/b46Aypc4 , https://pastebin.com/yZBmWTGL - idk what happened

[thegrb93]

physics crashes

[FunnyCraft]

what actions can this happen?

[Kefta]

The first is from vehicles, the second is from normal physics simulation. It's unknown why they occur as vphysics is not public.

[FunnyCraft]

The first is from vehicles, the second is from normal physics simulation. It's unknown why they occur as vphysics is not public.

https://github.com/Kefta/gs_crazyphysics maybe its help?

[Kefta]

They're not crazy physics crashes.

[GrahamBest]

Problem lies here: => 0xf3beb7f7 <+23>: movzwl (%ebx),%ecx

ecx is likely a nullptr which is why it's causing a segmentation fault. So whatever the first argument to that function is what's jacking it up. Seems to do something with: https://github.com/UAVXP/LeakNet/blob/master/ivp/ivp_collision/ivp_compact_ledge_solver.cxx#L274

It's calling convention is __fastcall, so first argument is ecx (and it's value is a nullptr, then it's being used by another function). Don't know much besides that, but it'll give a good start on what's directly causing the issue.

Are you potentially missing libraries on your server?

EDIT: Yeah, it seems to be apart of a different library than what is given inside of that repository. Are there any leads on it?

[thegrb93]

You're assuming that the hl2 beta leak of IVP source is a reliable representation of what's in gmod.

[GrahamBest]

You're assuming that the hl2 beta leak of IVP source is a reliable representation of what's in gmod.

It definitely is.... it's a library for collision. Not to mention Garry's Mod is built upon the engine HL2 uses. Source engine has underwent changes, but most of the time it's additions.

Also your point is irrelevant. If you actually read through the dump and go through the assembler for the specific function, and rebuild it in pseudocode, you'll practically notice it's near identical to the function I just linked...

I'm not responding to your pointless posts anymore. I'm not going to deal with the hissy fits you throw, just like the last one-sided conversation we had, which you ended up blatantly insulting me on after I proved my point.. more than once.

[thegrb93]

I spared you a lot more fallacies than that in your observation. I'm just pointing out weaknesses in your theories so that you can grow and stop being such a noob. Sorry if you think that makes you look dumb in the process, but learn to learn from mistakes.

a) Source2013 is miles ahead of beta hl2 b) It's not just a library for collision c) You speculated that 'its likely a null pointer' via the segfault when it could be any invalid pointer d) You are assuming the bug is in that function/stack-frame when fuck who knows where in the engine could have triggered such error state. e) missing libraries would not cause this, you #($&#$

EDIT: Yeah, it seems to be apart of a different library than what is given inside of that repository. Are there any leads on it?

That repository is beta hl2

[GrahamBest]

@thegrb93

1) Go read the source to each, not really. Major additions, yes. Worlds different? Not really. On top of it all that's completely irrelevant. The assembler clearly shows it's near, if not 1:1 identical to the function I provided.

2) Uh? No shit. I never specifically said that.

3) Uh, didn't I say likely? Get off your high horse and go read the assembly, you "noob." Of course it might not be a nullptr, I was simply giving an example. It's still a segmentation fault with what I linked.

4) Bumping number 3, if you had any idea how any of it works you wouldn't make such a clueless statement. The error is in the dump for Christs sake.

5) No, it definitely can if OP doesn't have the valid function the ECX register gets pushed on then called using. As said, if you open your eyes and go through the dump its pretty damn obvious.

You have proven you have absolutely zero idea what you are talking about, incredibly low self-control, an ego that depletes your learning abilities, and on top of it all, a failure to actually accept an answer that isn't your one line responses, by the way, more correct than yours. Stop spreading misinformation. Go read up on what I mentioned.

Okay, I made my point and proved you wrong.... again. I'm done having a kiddie argument of who's right. Now I'm seriously ending "conversating" with you. Have a nice day. I only wish you the best.

[Kefta]

Vphysics probably hasn't changed that much since 2003.

[thegrb93]

That could be true, but it doesn't change the fact this guy thinks he's the only one who can read a stack trace and signal and thinks he can solve the bug with his assembly skills.

@Verideth You're just oozing pretentiousness. Really, it was very impressive how you read that stack trace and where the signal was caught; very low level bug diagnosis. Oh and if anyone discredits any theory you shit out, they just need to read the link you googled and they'll understand what they did wrong. I honestly don't know what's more annoying: you thinking you proved me wrong, or you thinking you're contributing anything useful to these crash reports.

[thegrb93]

I get that crashes are probably what you're interested in, but this is not something you can solve. Unless you know the reliable way to reproduce the crash, then don't reply.

[thegrb93]

He was wrong on several accounts and this is the third time he's tried brushing off my criticism as if I'm some kiddie that doesn't know what I'm talking about.

[Kefta]

-snip- I wrote a vomited out a subpar analysis because I let my akrasia get the better of me. Don't step in the middle of an autistic food fight.

[GrahamBest]

@thegrb93 You are not giving criticism. You are completely throwing off my response as bullshit then proceeding to blatantly insult me or make a paragraph saying I'm wrong, with absolutely no reason why.

Now to respond to your most recent post of what you call "criticism." As @Kefta has repeatedly stated, what I said isn't wrong, and is in reality a rational and appropriate thought for the error. I never said my idea for the error was right. I do know you are incorrect with your statements, however.

[thegrb93]

@Kefta There's more ways to segfault than just null. Yeah, i left off the nitpicky criticism initially before he started brushing me off again. The whole point I'm making is he's not contributing anything useful to the conversation, at times is making wrong observations, and when you point it out he just tells you to fuck off.

[GrahamBest]

@thegrb93 Just stop. You have absolutely no idea what you are talking about and now have two people telling you otherwise, yet you fail to just listen.

[robotboy655]

Closing as stale since there's no info let in the original report. If the issue persists, open a new report.