> > Why is the angles check testing angles.x twice and angles.z > not at all? > > Shouldn't this just check angles.IsValid() instead? It's > probably not > > related to my issue, it's just curious. > > Maybe they're just checking to be doubly sure that angles.x > is valid. ;)
clearly angles.x is twice as important as angles.y. angles.z is not important at all. > > Nah, it's a bug in baseentity_shared.cpp in > CBaseEntity::VPhysicsUpdate(). It should be... > > if ( !IsFinite( angles.x ) || !IsFinite( angles.y ) || > !IsFinite( angles.z ) ) yes. > BTW, you definitely have a NaN problem in your angles.y value... > > "Invalid origin from vphysics! (4.000000,1.#QNAN0,-170.845169)" > > Maybe you have a divide by zero problem somewhere when > calculating the angles? It's in the origina though, not the angles. Usually this happens when vphysics gets bad input data. I'd put a similar .IsValid() trap in the places where you create physics objects e.g. do a search for "->CreatePolyObject" in src/game_shared and add an assert/trap to each call there. I've seen this come from unitialized memory in the past. Otherwise it could have been caused by some other pathological situation (applying velocity to a static object, setting a mass or inertia of zero, creating a constraint between two objects with NaNs in the constraint coordinate transform), but vphysics should be checking for these kinds of errors internally already. Jay _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
