Re: [hlcoders] Possible Crash

2008-12-22 Thread Justin Krenz
UTIL_Remove does exactly what you say. It removes the entity on the beginning of the next frame. It adds it to a list of things to delete, and that list is checked at a certain point of the frame when everything is supposed to be done with it. UTIL_RemoveImmediate is the bad one. It doesn't wai

Re: [hlcoders] Possible Crash

2008-12-22 Thread Michael Chang
I've had something similar happen once, although not directly associated with ragdolls. If I had to guess, it's something to do with using UTIL_Remove in an unsafe location in code, possibly during the an object's post-frame where something might be expected of an object after which it has been su

Re: [hlcoders] Possible Crash

2008-12-22 Thread Andrew Ritchie
The obvious question is, "Does it stop happening if you use UTIL_Remove?" I've not run into it, though our knock down code never hit public release and I used the SDK base, but I'm fairly sure I used UTIL_Remove and was ok other than the god aweful bandwidth hogging. On Mon, Dec 22, 2008 at 10:37

[hlcoders] Possible Crash

2008-12-22 Thread Justin Krenz
While working on my current mod (Smashball - www.smashball.tv), I came across a rare server crash in physics.cpp: line 454 (int CCollisionEvent::ShouldCollide function). This function tries to determine if two entities should collide. The physics system has the physics objects for each object alo