devtrackers.gg
Haha wow, didn't expect that. My guess is that the code kicking in when on a moving platform is responsible for this. I'll ask QA to reproduce it and look into a fix, thanks for sharing :)
Haha wow, didn't expect that. My guess is that the code kicking in when on a moving platform is responsible for this. I'll ask QA to reproduce it and look into a fix, thanks for sharing :)
This seems to be related to the video posted yesterday of the guy jumping from the bottom of point B devastation conquest over the barricaded glass doors. But landed on the 2nd floor where the mcom would be in rush.
Or the guy reviving through a wall.
Game just "loses" you for a second and by the time it catches up, it doesn't know you've passed a hard object. Waiting for people to use this to get OOB but still in the game, I mean it has to be possible if this is possible.
In the case of vaulting it makes sense since physics is getting bypassed for a while, but in this case there shouldn't be a valid reason for the physics body to go through the terrain. No matter the angle or velocity, physics won't let you pass unless something else goes wrong, so the question is: which system is responsible in this case.
What my uneducated eyes see is that regardless of physics and angle/velocity, he falls from 2-3m which doesn't seem like long enough for the game in this scenario to catch that hes actually hit a hard obstacle, so he just goes through like it never existed. Isnt this the same issue that plagues pretty much any online game? Didn't I use this same issue to get OOB in Tony Hawk PS as a kid? Travelling at a speed with latency? Having the game to try and relocate where I should be but having passed a solid barrier and placing me on the other side where I "was last" in a sense? But without the rubber banding actually being visible?
With a very basic physics system your observations would be valid, but with the one we use these things shouldn't matter. Moving objects in a physics system are not simply teleporting from point A to B. Instead there are many ways to detect intersection of moving objects. I don't know the specifics, thankfully I'm not a physics programmer, but you can think of 'sweep checks', where in a given time like 1/60 of a second (one frame) objects are compared based on their location and velocity. The system will be smart enough to detect very high velocities and act accordingly.
Now there are systems that override this behavior, like the vaulting system, or when you're attached to a vehicle for example. What happens here I don't know, but our QA will see if they can reproduce it which should at least point us in the right direction.
Ahh I see thanks for discussing your thoughts. Much appreciated. I see you guys are individually trying to help out here whenever you can lately so I'm sure its appreciated and not going unnoticed by me at least.
And if what I'm reading into is correct, our inputs in this physics based movement system aren't as much of the old "move to grid x=**** y=***** z=*****" but an actual velocity in which checks occur and could be performed ahead of time based on that velocity and what's possible. Which doesn't need to be present while hard animations/inputs like vehicle exits and vaulting occurs. But then picks up again once the animation and pathing is finished. If I were a physics programmer, I would think my first act in trying to reproduce it is to try and break that transition period between the two? Good luck on locating it. Have a good day.
Yeah exactly, someone would have to try to reproduce the same thing and then check all systems involved on what is currently going on. I'm surprised of it happening here since afaik there is nothing overriding physics, but that makes the issue more fun to fix :)